Weathering of Plastics: Testing to Mirror Life Performance (Plastics Design Library)

Published in the United States of America, Norwich, NY by Plastics Design Library a division of Willialn Andrew Inc. Information in this document is subject to change without notice and does not represent a commitment on the part of Plastics Design Library. No part of this document may be reproduced or transmitted in any fonn or by any means, electronic or mechanical, including photocopying, recording, or any information retrieval and storage system, for any purpose without the written permission of Plastics Design Library. Comments, criticism and suggestions are invited and should be forwarded to Plastics Design Library. Plastics Design Library and its logo are trademarks of William Andrew Inc.

Please Note: Great care is taken in the compilation and production of this volume, but it should be Inade clear that no warranties, express or implied, are given in connection with the accuracy or completeness ofthis publication, and no responsibility can be taken for any claims that may arise. In any individual case of application, the respective user must check the correctness by consulting other relevant sources of information. The use of general descriptive names, registered names, trademarks, etc. in this publication does not in1ply, even in the absence ofa specific statement, that such names are exempt fron1 the relevant protective laws and regulations and therefore free for general use.

Plastics Design Library, 13 Eaton Avenue, Norwich, NY 13815 Tel: 607/337-5000 Fax: 607/337-5090

Table of Contents Preface

George TVypych Basic Parameters in Weathering Studies George Wyp)Jch and T0111 Faulkner Choices in the Design of Outdoor Weathering Tests LarlY W Masters and Laurence F. Bond A Comparison of New and Established Accelerated Weathering Devices in Aging Studies of Polymeric Materials at Elevated Irradiance and Temperature Jorg Boxha111111er and Kurt P. Scott Current Status of Light and Weather Fastness Standards. New Equipn1ent Technologies, Operating Procedures and Application of Standard Reference Materials Jorg Boxhal11111er Weatherability of Vinyl and Other Plastics Jal11es W SUl11111ers and Elvira B. Rabinovitch Aging Conditions' Effect on UV Durability Robert L. Gray, Robert E. Lee, and Brent M Sanders Molecular Weight Loss and Chemical Changes in Copolyester Sheeting with Outdoor Exposure D. R. fa*gerburg and M E. Donelson Fourier Transform Infrared Micro Spectroscopy. Mapping Studies of Weathered PVC Capstock Type Formulations. II: Outdoor Weathering in Pennsylvania Dana Garcia and Janine Black Effects of Water Spray and Irradiance Level on Changes in Copolyester Sheeting with Xenon Arc Exposure D. R. fa*gerburg and M. E. Donelson Hot Water Resistance of Glass Fiber Reinforced Thermoplastics Takafu111i Kawaguchi, Hiroyuki Nishilnura,Fu111iaki Miwa, Kazunori Ito, Takashi Kuri}'a111a, and Ikuo Narisawa Surface Telnperatures and Temperature Measurement Techniques on the Level of Exposed Samples During Irradiation/Weathering in Equipment Jorg Boxhanl111er

Vll

61 69

93 99

105

Table of Contents

Infrared Welding of Thermoplastics: Characterization of Transmission Behavior of Eleven Thermoplastics Hong Jlln Yeh and Robert A Grimm Infrared Welding of Thermoplastics. Colored Pigments and Carbon Black Levels on Transmission of Infrared Radiation Robert A GriJnnl and Hong Yeh Predicting Maximum Field Service Temperatures From Solar Reflectance Measurements of Vinyl Henry K. Hardcastle III Residual Stress Distribution Modification Caused by Weathering Li Tong and J R White Residual Stress Development in Marine Coatings Under Simulated Service Conditions Gu Yan and J R White Balancing the Color and Physical Property Retention of Polyolefins Through the Use of High Performance Stabilizer Systems M J. Paterna, A. H Wagner and S. B. Salnuels Activation Energies of Polymer Degradation Sanlliel Ding, Michael T. K. Ling, Atul Khare and Lecon Woo Failure Progression and Mechanisms of Irradiated Polypropylenes and Other Medical Polymers L. Woo, Samuel Y Ding, Atul Khare, and Michael T. K. Ling Chemical Assessment of Automotive Clearcoat Weathering R. O. Carter III, John L. Gerlock and Cindy A. Sn1ith Effect of Aging on Mineral-Filled Nanocomposites A. Ya. Goldnlan, J. A. Montes, A. Barajas, G. Beall and D. D. Eisenhour The Influence of Degraded, Recycled PP on Incompatible Blends Claudia M. C. Bonelli, Agnes F. Martins, Eloisa B. Mano and Charles L. Beatty Interactions of Hindered Amine Stabilizers in Acidic and Alkaline Environments K. Keck-Antoine, D. Scharf and H. Koch Interactions of Pesticides and Stabilizers in PE Films for Agricultural Use Edina Epacher and Bela Pllkanszky The Influence of Co-Additive Interactions on Stabilizer Performance Robert L. Gray and Robert E. Lee

121

127

133 141

151

161 169

177 185 195

211

217 225 233

Table of Contents

New High Performance Light Stabilizer Systems for Molded-in Color TPOs: An Update Peter Solera and Gerald Capocci Stabilization of Polyolefins by Photoreactive Light Stabilizers Gilbert Ligner and Jan Malik Effect of Stabilizer on Photo-Degradation Depth Profile T. J. Turton and J. R. White New Light Stabilizer For Coextruded Polycarbonate Sheet Jalnes H. Botkin and Andre Schn1itter Ultraviolet Light Resistance of Vinyl Miniblinds. Part 2. Reaction Products Formed by Lead in Air Richard F. Grossnzan Case Studies of Inadvertent Interactions Between Polymers and Devices in Field Applications Joseph H. Groeger, Jeffrey D. Nicoll, Joyce M Riley, and Peter T. Wi·onski Automotive Clearcoats George W}pJ)ch and Fred Lee Index

241 253

261

271 277 281 291 315

PREFACE Before synthetic materials found a place in our lives, men and women relied on natural materials to build their houses, churches, buildings, to make their clothing and all other articles which societies required. These "traditional" materials were used with little or no chemical conversion. Natural forces determined which materials were durable and which were perishable. Our forebears learned by observing natural effects which n1aterials should be used for long-term use and which were disposable. At the end of their useful life, disposal of the articles caused little environmental impact as these natural products once again became part of nature. Today we have become engulfed with products and n1aterials made from materials extensively modified from their original, natural state. These modifications are often done in chemically irreversible ways. We want the products to be durable over their useful life but we also want theln to be returned to nature when we no longer need them. We hope that their disposal will not cause pollution. We need our water to be pure, our air to be safe to breath, and our soil to be uncontaminated. Conflicts abound. If we are to resolve them and continue to use synthetic materials responsibly, we must plan carefully and gain a complete understanding of how materials will perform and degrade. In particular we must be able to understand how materials weather, what the by-products of weathering are and how materials can be transformed into non-polluting entities either through recycling or natural disposal. Terms such as "life cycle assessment", "recyclable", "biodegradable" and "lifetime warranty" slip easily off our tongues. We need to bring weathering testing to the point at which reliable testing and investigative studies can enable us to use these and related terms with con1plete confidence. In spite of the efforts of research groups, standardization organizations and industry, there is much to be done to bring weathering testing to the level that will allow the results to predict the life of materials. There must be a willingness among the involved parties to cooperate and a con1prehensive body of information to support their efforts. This book is a contribution to the information base to assist the scientific efforts aimed at improving the knowledge of weathering. ChemTec Publishing and William Andrew will continue to supply infonnation to this field. In the year 2000 we will publish: • The 3rd Edition ofthe Handbook ofMaterial Weathering which will focus on information to support weathering testing • The Atlas ofMaterial Dalnage, a CD-ROM and on-line database of visual images characterizing various modes of degradation, their morphological features and reasons for the effects

viii

Preface

• The Manual of Testing - a collection of methods of testing used in various industries and research laboratories written by experts in their fields • Weather Data on CD-ROM - a collection of information on weather designed to assist experimenters in selecting the appropriate conditions for laboratory studies. One aim of this book is to provide a critical overview of methods and findings based on experimental work. Another is to create an awareness of the effect of the combined action of all the weather variables on materials under study. The introductory chapter outlines experimental design techniques and equipment selection and etnphasizes the importance of selecting the basic parameters of weathering including: • UV radiation • temperature of the specimens • rainfall and condensed moisture • humidity • pollutants • stress The book is structured to illustrate the importance of these parameters on weathering studies. Throughout the book, the authors attempt to show that weathering is not only dependent on UV radiation but that the overall effect depends on the interplay of all parameters which create a unique sequence of events that will change if the parameters are changed. The lack of correlation between laboratory and outdoor exposure is frequently caused by combinations of factors among which the improper selection of laboratory conditions is prime. After the introduction we discuss the choices available for outdoor weather testing. This relates laboratory tests to tests outdoors so that there may be correlation with natural conditions. The importance of precise control of both UV spectral intensity, temperature and heat flow is demonstrated in Boxhammer's careful use of available equipment and by studies done on automotive components. The recent availability of the ClRA filters and the continued use of borosilicate filters now permits accurate duplication of solar radiation. The chapter by Summers and Rabinovitch shows how radiation wavelength impacts the performance of several polymers. The manufacturers of weathering equipment can perfectly simulate the solar spectrum. Researchers now must take advantage of these developments. We show that failure to duplicate the solar spectrum invalidates the experiment. The failure is caused by energy input, temperature, moisture, and radiative effects. These parameters should not differ in the experiment from that of natural exposure. We compare the two most common artificial light sources - xenon arc and fluorescent lamps. The automotive, textile, polymer and stabilizer industries use xenon arc which gives the full spectrum of solar radiation (UV, visible, and near infrared). The use of fluorescent

Preface

lamps, which lack the spectral range ofthe xenon arc, should be discouraged except in special cases where the known mechanisms for degradation are triggered only by radiation between 295 nm to 350 nm. Several industries report problems stemming from studies done with fluorescent lamps which fail to correlate with actual outdoor exposure. Water spray during weathering studies has often been neglected. The reported work on co-polyester sheeting shows how complex material changes can be in the presence of water. More work is urgently needed to determine how hUlnidity and condensation influence material degradation. Two contributions from the Edison Welding Institute have been included to den10nstrate the effect of infrared energy and how different materials absorb this energy differently. In particular, the inclusion of pigments complicates infrared absorption. The chapter by Hardcastle shows how an evaluation of performance requirements helps to define a method of predicting the Inaximum allowable service temperature of vinyls based on measurements of their solar reflectance. Products in service operate under mechanical stress due both to residual stresses developed during the forn1ing process and to external stress in use. It has long been recognized that stress affects weathering but little has been done to evaluate the effect. Two chapters by White et ale propose methods of evaluating the effects of stress in weathering studies. These effects are complex since the initial stress distribution changes during exposure and this requires a knowledge of the kinetics of these changes. A similar situation exists with respect to the effects of pollutants. We know they influence weathering but there are few studies that assess their influence. Paterna et ale examine gas fading of automotive components in the presence of nitrous oxides. More elaborate techniques must be developed to evaluate the combined effects of UV radiation, moisture, temperature and pollutants on products to sin1ulate outdoor applications. It is unrealistic to study these influencing factors independently. Two studies on the effects of high energy radiation have been included to den10nstrate well defined projects which evaluated material failures and determined the activation energies of the degradation process for many materials, explained why degradation occurred in industrial sterilization, and determined how such degradation might be prevented. Assessment of automotive clearcoats and nanocomposites show that current test methods are sufficiently accurate, sensitive and suitable to detect degradation at an early stage of exposure. This is another area where more investigative work is needed. The benefit of this approach lies in gaining information early in the product developn1ent process using the equivalent of natural conditions without depending on the use of high energy radiation, often employed in accelerated testing, which causes degradation mechanisms which would not normally occur. Several contributors emphasize other complexities which must be dealt with in weathering studies. The materials themselves are complex. Many contain additives which interact with the host, the substrates and one another in a weathering situation. Conclusions may err if

Preface

they are based on an inaccurate knowledge of the real composition of the material under study. Even the manufacturer may be unaware of the true composition as composite additives may have proprietary compositions which are not disclosed. Many fundamental studies are needed to investigate the interactions of multi-component systems and to unravel the effects of processing aids which may be added without knowledge of their effects or interactions. Such practices may lead to unexpected and possibly, catastrophic, failures which would remain undetected in routine research and quality control operations. The stabilizer manufacturers have, as an industry, made a significant contribution to weathering testing methods. There are several chapters from these sources. They show that their reports to their customers are meticulous in relating the results of evaluations to the conditions of the test. Their approach is conservative in selecting both equipment and test conditions. The tests are expensive. They must relate to the real conditions of use and results should be comparable to those of prior tests. The book concludes with an example of the type of ground work and planning that is required before routine analysis begins. Using work on automotive clearcoats, we demonstrate how information must be analyzed and categorized to provide a rationale for testing, defining performance requirements, exposure conditions, mechanisms of degradation and how best to observe and measure the changes in specimens. Information gleaned from field performance is used to determine the appropriate laboratory simulations. If this preparatory work is not done the subsequent testing efforts are unlikely to yield useful data and be of little use in predicting future product performance. One final comment. Manufacturers must operate to meet economic goals. Industry as a whole is becoming increasingly competitive and is continually seeking ways to rationalize production methods to improve econonlics. Materials from different industries compete for the same markets. Durability has become one ofthe most important characteristics. The product is either made from an inherently durable material or it receives an external coating which gives the required durability. The first approach is more consistent with recycling processes which generally have difficulty in dealing with multi-component mixtures. As the understanding of weathering increases we may learn how to more frequently select a durable substrate which will not require the complication and cost (initial and recycling) of a surface coating. The economic answer would seem to lie in making the investment in weathering research to avoid the costs of material replacement and material failures. I sincerely hope that this introductory volume will generate an increased interest in advancing these important studies and provide an inspiration to researchers to pursue weathering studies as both economically and environmentally important activities. George Wypych ChelnTec Laboratories, Inc. Toronto, Septelnber 1999

Basic Parameters in Weathering Studies

George Wypych Che111Tec Laboratories Inc., 38 Earswick D,:, Toronto, Ontario M1E 1C6, Canada Tom Faulkner Atlas Electric Devices Company, 4114 North Ravenswood Ave., Chicago, IL 60613, USA

INTRODUCTION In spite of the efforts by manufacturers to produce durable goods materials do fail. These failures not only affect custon1er perception of the abilities of manufacturers to deliver products designed for the required perfonnance, but also result in complaints and liabilities. We know from everyday practice that products do fail and examples such as paint peeling from cars, faded and discolored textiles and plastics, or various defective construction materials are con1ll1onplace. Many of these failures are caused by the exposure of materials to the environmental conditions which include factors listed in Table 1. Two observations from this table are important: • degradation rate is controlled by a set ofparameters that can affect results oftesting • the material testing in real environment is affected by the variability of weather The discussion of weather conditions in various parts of the world and also in one location shows very large variations, such as, seasonal, geographic and weather variations from year to year (see reference 1). These variations n1ake testing in the natural conditions very difficult because only long-tenn testing results can average these variations in climatic conditions and thus results. This is one reason underlining the need to test materials in a laboratory under conditions which can always be repeated. It is known from any type of study that if parameters of an experiment are not strictly controlled the results of study are meaningless. This, in turn, shows the need to choose adequate equipment and select proper parameters of testing. These subjects are discussed below. There is also a need in the studies on the n1aterial durability to select a yardstick which can be used to obtain results in a numerical form pennitting comparison of the results. Here, two matters are important: method of specimen testing and reference standard to which these testing results are related. The methods of specimen

Weathering of Plastics

Table 1. Parameters of material degradation Parameter

Typical range

Comments

UV radiation

295 to 380 nm

UV radiation in this range is found in the sun radiation. UV radiation below 295 nm causes degradation that does not occur in real life

Air temperature

-40 to 40°C

Air temperature is rarely the same as the product temperature because products also absorb infrared radiation

Product temperature

-40 to 110°C

Actual product temperature is a composite of air temperature, effect of infrared radiation, effect of wind, and surface evaporation of water. Product temperature is a parameter which must be selected for testing

Rain

o to 2500

Rain is important because it washes away components of material and deposits dissolved gases such as carbon dioxide, oxygen and pollutants (e.g., acid rain)

nlm/year Relative humidity

10 to 100%

The relative humidity participates in degrading some components of the materials and in deposition of pollutants

Pollutants

variable

Pollutants include carbon oxides, ozone, oxides of sulfur and nitrogen, radicals, dust particles. These pollutants can be deposited by rain to become more aggressive degradants

Stress

variable

Materials degrade more rapidly under the mechanical stress

testing are discussed below in a separate section. The reference standard of laboratory results is the material perfonnance under its nonnal conditions of use. This brings us back to the exposure to environmental conditions. The choices of selection of exposure sites and the conditions ofsuch exposure are omitted in this discussion. But, it should be borne in mind that the results of long-term testing of the same or similar materials in the weathering stations allow us to express the results of laboratory studies in the required form of years of product performance by correlating them with results of laboratory studies. Planning durability testing of a material requires not only proper strategies to chose adequate methods of testing and exposure but reasons for testing should also be evaluated. It is quite obvious that user of material requires durable product but the use of material is complicated by some additional considerations such as • design life period of use intended by designer of structure in which material perfonns its functions • dealing with failure replacement, maintenance, lifelong • cost of replacement nlaterials, removal, installation, disposal • post application plans recycling, disposal, renewable resources

Basic Parameters in Weathering Studies

These additional criteria must be factored in the entire plan regarding use of many materials. They decide about effectiveness of material use and associated costs. At the same time, the importance of these factors puts even more stringent requirement on the quality of testing results. Having in mind that the results of testing affect decision making process of product selection and its economy of use, one must conduct these studies in a manner that gives assurance that the outcome of testing gives reliable information on product behavior in real life. This introductory chapter gives a general overview of selection of testing conditions. This information is further elaborated in other pal1s of the book.

SELECTION OF PARAMETERS OF EXPOSURE In this section we will analyze further the choices of parameters listed in Table 1 and the potential implications of choices on the predictive value of the testing results. UV radiation is, for most materials, the most important determinant of their durability and as such deserves considerable attention. Two factors help to quantify UV radiation: solar cut-on wavelength and itTadiance. The solar cut-on wavelength is the lowest wavelength still available in the sun radiation. The value of a solar cut-on varies with the season and it is commonly estimated at 295 run in sununer and 310 nm in winter. Below these cut-on values there is no radiation in daylight. Considering that the lower the wavelength, the higher the energy of radiation, the sun radiation is less damaging to material in winter than in summer. What does happen if we perform the tests using radiation of a lower wavelength (e.g., 260 run)? It can be expected that, since radiation at 260 run has higher energy than at 295 run, the damage of material should be more extensive because more radiant energy was applied. This faster degradation is not, by itself, precluding the lower radiation wavelength from use because we want to obtain test results faster. But, other question arises. Is the degradation process the same when we use radiation of a higher energy? The answer is no. There are two reasons for this: materials have selective absorption and reactions occur exclusively at certain energy levels. The selective absorption means that any given material is capable of absorbing only at certain wavelengths (but not at the others). These bands of absorption are the characteristic properties of any given material. For example, polycarbonate exposed to three wavelengths ofradiation 260,280, and 300 run degraded extensively at 280 nn1 because it does not absorb radiation at 260 and 300. Thus, radiation at 260 run, having higher energy than radiation at 280 run, was harmless because energy is used for degradation only when absorbed by the material. On the other hand, if polycarbonate was exposed to radiation fron1 a lamp which had UV radiation in a range fron1 260 to 380 (such as for example mercury lamp), polycarbonate would show signs of degradation because it absorbs radiation at 280 nn1 which does not exist in daylight (the energy of

Weathering of Plastics

radiation at a wavelength of300 run, which is present in daylight, is not sufficient to degrade bonds in polycarbonate.)

The conclusion from the above is that no radiation below the solar cut-on (295 nm) should be present in the equipment used for testing. Many other examples of real n1aterials support this statement. Irradiance level is the second factor which determines energy of radiation. Irradiance is the rate with which sun or lamp energy falls on the surface ofn1aterial. It is expressed in Watts (units of energy) per surface area (usually m 2) and a wavelength. From this definition one may expect that the more energy falls on an object, the more damage can be expected (providing that energy is absorbed). Two characteristic values of irradiance are used in practice. These are 0.35 and 0.69 W/m 2 at 340 run. The value 0.35 W/m 2 is natural daylight irradiance measured at 340 nm at 26° in Florida and 0.69 W/m2 is a peak value of natural daylight irradiance. If irradiance in laboratory testing is above these values, test results may not be comparable with the results of exposure to natural environmental conditions. The use of higher irradiance values requires additional studies which prove good correlation between laboratory and natural exposures. In conclusion, irradiance setting at 0.35 W/m 2 at 340 run should be used for most laboratory testing to obtain reliable data. The operation of an instrument under high energy levels speeds up the process ofdegradation but the results ofstudies may not reflect the performance characteristics of materials used under normal conditions. In the last section, it will be shown that safe methods exist which allow for early detection of failure (or acceleration of testing). The selection of higher irradiance usually requires that the preliminary experiments confirm that the mechanisms of material degradation were not affected by the high energy levels used for testing. Air and product telnperature. Temperature of a sample during testing has impact on results. Typical samples tested in a laboratory have different colors. Therefore, they have different ability to absorb infrared energy. Figure 1 shows the difference in temperature between white and orange colored samples. Temperature depends on time of the day and color. It is known from practice that black specimen may reach 90°C which is about 30°C higher than for a white specimen under the same conditions. The highest temperatures ofup to 110°C where recorded inside the enclosed cars. The rule ofa thumb in chemistry assumes that the reaction rate doubles with every 10°C increase of temperature. Therefore, black sample, according to this rule, should be degrading 8 times faster than the white sample. It can be concluded that samples should be tested in their real temperatures, resulting from ambient air temperature, absorbed infrared energy, and cooling effect of water evaporation.

Basic Parameters in Weathering Studies

60 Temperature, deg C orange 50

20 L 11

- - - - - - - - - - -....

Daytime, h Figure 1. Temperature behind the sample on sunny day. [Adapted froln G. Wypych, Handbook of Material \Veathering, Che11lTec Publishing, 1995].

Rain and relative hll111idity. The analysis of typical weather conditions shows that rain occurs in 10-16% of the average day time in Central Europe. The amount of rain varies even more widely when one COlnpares dry climates (e.g., Arizona) with subtropical climate (Florida). In Miami, Florida, surface ofn1aterial is wet by average for close to 50% ofthe tin1e ofa year whereas in Wittmann, Arizona by average for about 4% of the time ofa year. Typical condition of operation of laboratory devices is 15% of the rain time. This parameter is essential because many additives in plastics operate on the surface of materials. The most typical additives include UV stabilizers, antistatics, and biocides. There are also numerous other essential additives such as, for example, plasticizers which tend to migrate to the surface to equilibrate for the lost concentration. If excessive rain or condensation is selected then n1aterial loses its properties without correspondence to natural conditions. Similar results are due to condensation if excessive humidity is used. Some polymers are also affected by moisture. Polymers such as polycarbonate, polyester, polyamide and many others hydrolyze in the presence of water. The hydrolysis is time-related and water concentration-related phenomenon therefore the increase in water supply to the sample changes mechanism of degradation. Typical settings of relative hun1idity are in a range from 30 to 90% with 50% used most frequently during a light cycle. Excessive condensation and selection of excessive rain changes the mechanism of degradation making results of laboratory studies not comparable with the nonnal conditions of performance of materials. Pollutants. Simulation of the pollutants influence is difficult to conduct in the laboratory equipment because of many different substances involved and their highly variable concen-

Weathering of Plastics

tration. The combination with typical weathering studies is complicated by the fact that specialized equipment is needed for such studies operated under variable conditions and variable compositions of pollutants. For this reason these studies are not a part of main stream weathering studies. Stress is an essential parameter of weathering considering that thermal and moisture movements in materials are found in practical applications and are known to affect the rates of degradation. Two aspects of stress interference can be considered: effect of external forces and residual stress retained in nlaterials after processing. The existing sample holders allow to induce static stress to material exposed to radiation and other environmental conditions. This mode of testing is one of the methods to accelerate testing and frequently to obtain results which are common with material performance in normal applications. The stress applied should be selected based on the prior knowledge ofmaterial performance conditions. The scientific literature makes suggestions to test specimens at 30 to 600/0 of elongation at break. At the same time, it should be considered that stress is an additional parameter of weathering therefore its introduction changes both rate and mechanisms of degradation. For this reasons, the effect of stress should be tested on well defined specimens. Based on the above discussion the suggested choices of main parameters are summarized in Table 2. This check list is useful in evaluation of laboratory equipment which can be used for testing giving high correlation with natural conditions.

Table 2. The list of important considerations for the selection of laboratory equipment which performs testing under conditions that may give a high correlation with natural exposures in addition to full spectrum of visible and IR, the UV radiation wavelength is limited to 295 to 380 nm irradiance can be selected and controlled within the range

0.35 to 0.8 W1m2 at 340 nm

sample temperature (not air temperature) can be selected and controlled by black panel 25 to 110°C and ambient air control within the range the sample temperature depends on color

temperature of samples varies

the real temperature of a specimen can be measured water quality can be controlled for water solids

<1 ppm

the rain duration can be selected and applied in selected intervals

o to 50% of experiment time

the relative humidity can be selected and controlled

30-100%

Basic Parameters in Weathering Studies

SELECTION OF EQUIPMENT FOR THE LABORATORY STUDIES Xlnan with elM Innl' Ind Sada Lime Outer Filters VI. Mllm' P.lt Sunlight, 28' South &POlUN

4....---------------.. . .-. -Sunlight

- CIRA/8Oda Ume

o .a-_.,....__~ - -.....--..._--....- _ - I

Wavelength (nanometres) Figure 2. Spectrum of a xenon lamp with ClRA/soda lime conlpared with daylight radiation. [Courtesy of Atlas Electric Devices Company, Chicago, USA]

UV Cut-off af Valiaus Flllir ComblnatlDns (0.56 W~ @340nm) VI. Miami Sunlight. 28' South Exposure 1.2&

Wavelength (nanometrea)

Figure 3. Effect of different filters on a light spectrum of a xenon lamp. [Courtesy of Atlas Electric Devices Company, Chicago, USA]

This discussion is focused on the major types of the equipment used for the weathering studies leaving aside the construction details related to a size ofequipment and its level of computerization which can be found from the equipnlent suppliers. This narrow focus is selected to concentrate on the capabilities of equipment to give reliable data which correlate with normal conditions under which products perform. The equipment technology b b I .fi d dO can e est c assl -Ie accor lng to the types of energy sources used by these equipments. Three major sources of radiation are used: carbon arc, xenon arc, and fluorescent lamps. The source of radiation determines the conditions of sample exposure. A carbon arc is an older technology which is still in use for testing samples according to standards developed for this equipment. The majority of standards either specify testing equipment with a xenon lamp, a fluorescent lamp, or give the choice of both. In

many existing standards, the equipment using a fluorescent lamp is given as an optional choice for fast or preliminary screening of samples because it is less costly and easier to maintain but also has numerous limitations. In this section, a COlnpar-

Weathering of Plastics

8 UVA·340 and F40 UVB Fluorescent Sunlamps vs. MlamlllPeaku Daylight 2.00 - r - - - - - - - - - - - - - - - - - - - , 1.60 1.20

0.80 0.40

300

400

roo

800

Wave1ength (nanometres)

Figure 4. Spectrum of fluorescent lamps compared with daylight radiation. [Courtesy of Atlas Electric Devices Company, Chicago, USA]

EEl II

Orange

BI ue

outdoor xenon arc fluorescent lamp

f:t±::t::t::t::t::t:t::t::t::t::t::Jt::tjt:::t:i=t::t::t:tt

j;t::t:::t:t::t:t:±::t::t:::::t::t::t::::jt:::t::tt:±::t::t:t:±t:tt:t:j

10 20

30 40 50 60 70 Temperature,

°c

Figure 5. COlnparison of temperature of colored samples. [Courtesy of Atlas Electric Devices Company, Chicago, USA]

ison of both types of equipment is made to show potential tradeoffs related to the application of either technology. Figure 2 shows spectra of daylight and a xenon lamp equipped with filters. It is apparent that the daylight radiation from sun is very well simulated in all regions (UV, visible, and also infrared range not shown on the graph).2 Figure 3 shows the effect of different filters on UV cut-on points. Filters such as ClRA and borosilicate glass provide radiation restricted to the range available in daylight. 2 Figure 4 shows spectra of a daylight and fluorescent lamp. The fluorescent lamp of UVA-340 type can match sun daylight between 295 and 350 run. Outside this range the lamp does not emit substantial levels of radiation. This characteristic has implications on the results. First of all many materials absorb and degrade outside this range. Polyamide-6 (nylon) is one such example since it has substantial degradation when exposed to the radiation wavelength at 365 run. Pigments are known to change colors in the visible range producing products of degradation which affect stability of resins. Use of such pigment may cause

Basic Parameters in Weathering Studies

Ugh: . .

'Janitor

Figure 6. Schematic diagram ofWeather-Ometer. [Courtesy of Atlas Electric Devices Company, Chicago, USA]

degradation of material in field which was not detected in a laboratory performed in the fluorescent device. There are numerous other examples which can be found elsewhere. 3 Figure 5 illustrates other important problem related to temperature ofdegradation. Samples exposed to an outdoor conditions and xenon lamp are degraded under different temperatures depending on their color. The temperature of samples in a fluorescent light device can only be controlled by heating the chamber but this results in the same temperature for all the samples which differs from their exposure in normal conditions of their performance. Many confusing results were produced because of this discrepancy. 2

The third important problem is related to the handling capabilities of humidity and rain. The fluorescent devices do not have control over humidity. Their controls can be in one oftwo positions: on or off, and rain is an optional accessory. Water is delivered to samples by evaporation from reservoir and condensation of vapor on the surface of samples. This increases probability of washing away some vital components as was previously discussed. Also, conditions of water deposition differ from natural environment in which rain is a factor (not available here) and condensation occurs at different temperature range (outdoor exposed material has temperature lower than ambient air, usually below 20°C, which causes moisture condensation.) These three deficiencies combined with many other simplifications introduced to decrease the cost of construction are behind the designation of the machine as suitable for preliminary screening of samples. The preliminary screening results should always be compared with the results obtained from devices equipped with a xenon lamp in fully controlled equipment as discussed below or long-term weathering studies. For reliable testing, equip-

Weathering of Plastics

ment giving more operational controls and having environment similar to outdoors should be used. Figure 6 shows a schematic diagram ofWeather-Ometer which is equipped with a xenon lamp. The instrument provides full control ofUV radiation, air temperature control, rain and rain water control. All essential parameters can be selected by manual controls or from the computerized interfaces. Input and output data can be acquired to the remote computer and automatically controlled. These instruments are equipped with diagnostic functions which help users in troubleshooting and returning equipment to normal operation mode. 4 The selection of an instrument requires consideration of many additional factors such as the expected amount of samples, duration of exposure, sample sizes and shapes, method of sample mounting, etc. Suitable equipment which may satisfy any practical need is available. The instruments manufactured range from small benchtop instrunlents to large chanlbers which can accommodate large parts ofmachines (e.g., full automotive bumpers) and even entire cars. At the same time, benchtop instruments have limited controls over the environmental conditions under which experiment is conducted. The user should be aware of this limitations which may affect results of studies. In conclusion, selection of equipment for required precision of testing is a complicated process, which can be narrowed down to the analysis of elements essential to obtain results predicting material performance in natural conditions of their perfOTI11anCe as given in Table 2.

EXPERIMENT DESIGN The selection of parameters of exposure and suitable equipment as discussed above are very essential steps in the way to obtaining meaningful results. But still there is a high potential for studies to fail ifexperiment is not designed properly. We now outline the essential elements of the experiment design which serve a broad range of applications of weathering methods. British Standard BS 7543: 1992 5 gives clear guidance to how such studies should be properly conducted. Here are the most important points. The following major tasks must be performed: 1 Identify the performance criteria 2 Identify possible degradation mechanisms 3 Identify important degrading parameters 4 Identify the range of changes of performance criteria 5 Postulate on how degradation can be induced in accelerated weathering tests 6 Select methods of measurement 7 Design and execute preliminary study 8 Conlpare results with long-term data and degradation mechanisms

Basic Parameters in Weathering Studies

Conduct the routine testing These tasks look complicated but some of the action items can be avoided by adequate selection of parameters of exposure and instrument selection. From the principle of similarity it is known that if conditions of degradation (UV, temperature, humidity, rain, etc.) are selected very close to the average conditions encountered outdoors then there is a limited probability that the changes occurring in the material are caused by different mechanisms than the changes in the outdoor exposure. The selection of equipment and its operating parameters, similar to encountered in real life, allows to omit tasks listed in points 3,4,5, 7, 8. This limits our tasks to a much simplified process. Identify the perfornlance criteria. This first stage means that certain parameters ofproduct performance should be selected together with a time scale for their occurrence. Suppose that product should not yellow, crack, and lose tensile strength. These are the most essential parameters for this product performance. Now, it is necessary to attach time values to these parameters. This time scale is constructed based on the product role in the marketplace. Products are divided by the length of time of service. For example, in construction industry, products are divided into five categories: temporary (10 years life-time or less), short life (10), medium life (30), normal life (60), and long life (120). The other parameters determining the time scale of performance are related to how this product is used in practice. Products can be replaced, maintained, or designed for a long life without maintenance. In the first and the last case there is no cost associated with the product maintenance but only life titne and cost of replacement differ. These two elements of cost (replacement or maintenance) must be factored into product cost and thus reflected in the choice of time scale. The result of this task is a table of parameters of degradation (yellowing, cracking, etc.) versus the time to their first appearance. Identify possible degradation 111echanis111s. While the prediction of a mechanism is not con1pulsory in the experiment designed based on the similarity of conditions (average UV, temperature, humidity, etc.) because there is no check ifmechanisms of changes outdoors and in a laboratory are the same. At the same time, it is still good practice to do this analysis (at least once) because it n1ay help to improve products based on the understanding which properties fail and why. In order to perform this task one needs to use literature data and/or own studies to analyze what causes each mode of failure (e.g., cracking, tensile loss, etc.) and outline chemical mechanisn1s of these changes. Identify inlportant degrading paranleters. The identification of degrading parameters is not required if conditions in laboratory and outdoors are similar because all parameters work in combination. The only help from such analysis may be that it allows to predict methods of stabilization to improve products. Identify the range ofchanges ofpeljormance criteria. The range of changes of performance criteria is analyzed as an additional check point if conditions in laboratory and

Weathering of Plastics

outdoors differ. By doing this analysis we want to find out if measured changes are not outside the normally observed range of changes in normal use of a product. In the controlled experiment the only significance of such analysis would be to establish frequency of testing and sensitivity of testing method but otherwise such analysis is not required. Postulate on how degradation can be increased in accelerated weathering tests. Because degradation occurs under very similar conditions it can be expected that an accelerated test induces the same changes as average climatic conditions should do. Thus, the identified mechanisms of degradation in normal use are the same as mechanisms active in laboratory study which only is better controlled and run under repeatable conditions. Select methods oflneasurelnent. This step is very essential. The most frequently asked question in accelerated test regards the degree or rate of acceleration. One source of acceleration is operation of equipment set at Miami conditions. This gives different acceleration factors for different geographic locations which may typically vary in a range from 2.5 to 10. But, the acceleration depends much more on the method of observation (testing). Conducting experiment under well controlled laboratory conditions gives us samples suitable for many sophisticated methods of analysis. If a sample is exposed outdoors, it does pickup dust particles which, residing on the surface, make sample unsuitable for many chemical analytical techniques which require clean surface for analysis. These methods will only analyze dust. A simple example of acceleration can be given based on previous experiences. A material durability was measured by the time when cracks appear on the surface. When this material was exposed to outdoor conditions in Toronto, cracks were observed after 2 years of exposure by a simple visual inspection. Exposure in Weather-Ometer operated under conditions listed in Table 2 with in adiance of 0.35 W/m2 induced cracks which could be observed by a naked eye after about 2000 h of exposure which gives acceleration by a factor of8.76. The same specimens observed under stereo microscope allow to detect cracks after about 400 h of exposure which gives an acceleration factor of 43 .8. If analysis of these samples was performed under SEM microscope enlarging cracks could be detected after 40 h of exposure in Weather-Ometer which gives an acceleration factor of438. There are many similar analytical techniques available for every type of analysis required to attain similar accelerations. The importance of the method selection cannot be overemphasized. It is not only an acceleration factor which is important but also the method should be selected based on realistic criteria related to field observations of the product and available techniques which can detect and quantify these observations. It is always advisable to use methods which lead to answer in a shortest possible way. For example, if tensile strength is an important parameter indicative ofor important for its performance there is no need to look for chemical analysis which correlates with tensile strength but simply tensile strength should be determined unless one goal of 4

Basic Parameters in Weathering Studies

studies is to detennine the mechanism of degradation for which chemical analysis gives supportive data. Design and execute prelilninaly studies. Preliminary study and evaluation of its results are only critical when the conditions of exposure are different from outdoor environment. Such experin1ent under well controlled conditions of exposure is only valuable if experimentation is made with various methods of chemical analysis and it is uncertain whether the method of testing is the right choice. Otherwise, it is possible to go directly to the analysis of material and COlnpare the data from laboratory studies with outdoor exposure to built predictive criteria and evidence for life predictions of material. Such data are invaluable because they can be used in future for evaluation of reformulated products based on laboratory studies alone. There are other numerous factors related to weathering studies which cannot all be mentioned in one short paper. Many answers can be found in the specialized monograph on the subject. 3 One chapter of this book6 gives an example of preliminary data generation for testing a product based on exan1ple of automotive coatings. In conclusion, the effort was made here to show that simple approach to studies of weathering is possible. This approach allows to concentrate on the essence of weathering studies which are meant to predict properties and durability of materials in the most efficient fashion. These studies are invaluable for both consumer and manufacturer alike. If conducted in a well planned experiment they will help in extending life of materials and prevent unnecessary and unexpected failures. If attention is focused on the modification of parameters existing in the natural environment these studies will most likely fail in spite of the work done because common physical and chemical principles have been violated.

REFERENCES 1. 2. 3. 4. 5. 6.

L. W. Masters and L. F. Bond, Choices in the design of outdoor weathering tests. This book. Ci4000 Weather-On1eter. Atlas Electric Devices Company, Chicago, USA, 1997. G. Wypych, Handbook of l\laterial ''''eathering, 2 nd Edition, ChemTec Publishing, Toronto, 1995. Ci5000 Weather-Onleter. Atlas Electric Devices Company, Chicago, USA, 1997. British Standard. BS 7543: 1992. Durability of building elements, products, and components. G. Wypych and F. Lee, Essential paratneters of degradation of automotive coatings. This book.

Choices in the Design of Outdoor Weathering Tests

Larry W. Masters and Laurence F. Bond Atlas Weathering Services Group, Miallzi, FL

INTRODUCTION Demands for new and more durable products have led to increased interest in data on weatherability, durability and service life ofmaterials. As a result, manufacturers and users need reliable weathering data on a wide range of materials and systems to aid in product development, materials selection, quality assurance, life cycle costing and warranty considerations. Studies carried out through international organizations (Jernberg et al-1997; Masters-1986; Sjostrom-1990 and 1991) and within organizations of specific industries such as construction or automotive (Bauer-1997; Bourke and Davies-1996; Fischer and Ketola-1994; Martin et al-1994; Misev et al-1999, Nichols and Darr-1998, Shirayama-1993, Wicks et al-1999) illustrate the increasing interest in durability performance of materials and systems, as well as the use of new and innovative approaches to life prediction. Accelerated tests, for example, have been widely used to shorten test times (Wootton-1996) and have included test chatnbers large enough to accommodate full scale automobiles (Severon-1998). In addition, Sjostrom's work points out increasing interest in drawing upon databases of actual in-service performance of buildings as a source of data to aid service life predictions. But even with the rapid technological improvements in life prediction techniques, outdoor or natural exposures, accelerated weathering machines, evaluation techniques and database development there is often mystery, mistrust, and misunderstanding of weathering and durability data.

APPROACHES TO WEATHERING Three major approaches are used in weathering tests: 1) outdoor (or natural) exposure, 2) chamber exposure and 3) accelerated outdoor exposure. The general principle of assessing the effect of weathering is to measure specific properties or performance attributes prior to

Weathering of Plastics

and after weathering exposure. The difference provides a measure of the change in the materials or systems caused by the weathering exposure. Outdoor exposure methods include both direct and under-glass exposure; the chamber tests include Carbon-arc, Xenon-arc, metal halide and fluorescent exposure; the accelerated outdoor exposures include Fresnel-reflecting mirror machines and various other exposure devices aimed at simulating in-use conditions.

CAUSES OF DEGRADATION Materials and systems exposed to various elements of weather typically are degraded most often by 1) sunlight, particularly the ultraviolet (UV) (295-385 nm) portion ofthe solar spectrum, 2) thermal effects, including high and low temperature as well as temperature cycling, and 3) moisture. In addition, other factors such as air contaminants, oxygen, salt, etc. can also contribute to degradation for specific materials. The parameters of weathering are variable. Thus, to better interpret results of weathering exposure tests and to compare results obtained from different sites or at different times, it is essential to measure the key weathering parameters. Atlas Weathering Services G-roup (AWSG), for example, measures and reports data on total and UV radiation at various test angles, air temperature, black and white panel temperature, time of wetness, rainfall, etc.

KEY CHOICES IN THE DESIGN OF OUTDOOR (OR NATURAL) WEATHERING TESTS SELECTING THE OUTDOOR OR NATURAL EXPOSURE LOCATION(S) Figure 1 (Masters and Wolfe-1974) is a map of the contiguous United States showing one author's classification of climate zones. It is obvious that many different clinlate zones are observed in the US; on a worldwide basis, even more zones can be identified. With the worldwide marketing of many manufacturers, it is important to have weathering data which pertains to as many of the climate zones as possible. For that reason, AWSG has established a Worldwide Network of 19 outdoor exposure facilities which, in addition to ten locations in the United States, include sites in France, The Netherlands, Saudi Arabia, Russia, Singapore, Canada, Japan, and Australia. However, most organizations which have weathering tests performed desire to limit the number of exposure sites used to reduce costs of testing. Traditionally, the plastics and coatings industries have addressed the above tradeoff by heavily using two extreme climates for outdoor or natural testing: a hot, dry climate (such as the Arizona desert) and a hot, wet climate (such as subtropical Florida). Thus, US commercial exposure facilities have focused heavily upon Arizona and Florida and industry has assumed that

Outdoor Weathering Tests

- ..

, _ ..

,.", >

.-, ........

•••

".~: • •

~ '.

« ..

I: . ~,:~~.' w

_.1 . . ."*. (I

...

Figure I. Climates of the world.

Table 1. Climatological data - various locations Location

Lat./Lon.

Average rain mm

Average humidity %RH

Annual solar energy total/total UV, MJ/m 2 80001330

Phoenix, AZ

33°54'N/112° 8'W

254

Miami,FL

25°52'N/80027'W

Sanary, France

43°8'N/5°49'E

79 64

Choshi, Japan

35°43 'N/140045'E

2010 1200 1825

6500/280 5500/285 4000/220

Lochem, Netherlands

52°30'N/6°30' E

715

37001190

data from these two harsh environments provide "worst-case" scenarios. Most often, test results from the subtropical area of South Florida are used as the weathering standard for building and construction materials and in the automotive industry. In addition to the hot, dry and hot, wet extremes, other frequently used climate zones for weathering are I) seashore (salt air), 2) industrial, 3) temperate and 4) freeze-thaw. Often, the controlling specification or test method defines the exposure conditions.

Weathering of Plastics

Figure 2. Aerial photograph of AWSG's South Florida Test Service, Miami, Florida Weathering Station.

A frequently asked question pertains to the extrapolation of data obtained from one climate to other climates. One means of addressing this question is to compare specific climatological data, such as is done in Table 1, from various sites. For example, materials subject to degradation by radiant energy would be expected to degrade much more rapidly in climates with high radiant energy (particularly high UV radiation). Another means of comparison is to use results from specific materials exposed in various climates. For example, after investigating thirteen sites representing a wide variety of global climates, Bores (Bores-1976) reported that the degradation rate of the coatings tested were dependent on the climate; and that reversal in ranking could occur depending upon the formulation. In other studies from Europe (Helmen and Hess-1979), it has been stated that, "I year outdoor exposure in Basel (Central Europe) is equivalent to approximately 6 months in Florida." Koch (Koch-1980) has indicated that normally one year in Florida is equivalent to two years in central Europe. Koch went on to state that when the acceleration factor due to 45° South exposure is considered, one year in Florida is equal to four years in Central Europe at 90° South.

Outdoor Weathering Tests

Southern Florida is cited as a primary location for real time 8000 , - - - - - - - - - - - - - - - - - , 100 exposure testing in the specifica..E 6000 tions of major industries. A ...., panoramic view of one of the ~:::J 4000 two South Florida Test Service o :x: 2000 (SFTS) exposure fields in Miami, FL is shown in Figure 2. The reason for the extensive use of southern Florida is the relatively stable year to year climate _ Avg. High Ambient Air _ Total 26° Direct which offers high amounts of __ Avg. %RH. • •••• WetTime - Hrs. sunlight, about 3000 hours; warm year round temperatures Figure 3. Climatic conditions in Miami, Florida. averaging 24°C (75°F); high huPHOENIX ENVIRONMENT midity, 79% RH; and rainfall amounts of 2010 mm (79 in) 10000 , . . . - - - - - - - - - - - - - - , 100 coupled with 4200 hours of total 80 8000 in wet time. (See Figure 3). 60 m ~ 6000 The central Arizona desert 40 Q has similar regular climate and ~5 4000 "* 20 ~ has become a worldwide recog- - - '-'--'- - --_ _:x: 2000 nized standard exposure o o ................................ environment for testing as well. Sunshine hours average some 4000 hours while average tem_ Avg. High Ambient Air Total 34° Direct peratures are 22°C (72°F) with ••••• Wet Time - Hrs. - - Avg. %RH summer daytime highs above Figure 4. Climatic conditions in Phoenix, Arizona 40°C (l04°F) and winter nighttime lows falling below O°C (32°F). The air is dry with an average RH of 32%, with rainfall averaging 254 mm (lOin) and total wet time as low as 375 hours. (See Figure 4). In addition to the hot, wet and hot, dry climates often used for materials weathering studies, materials are also often exposed in climates with cold winters and large temperature cycles. AWSG exposure facilities are available in Canada and Chicago, IL for such exposures and a number of research laboratories have exposure facilities in cold climates. MIAMI ENVIRONMENT

CO<

:€

Weathering of Plastics

Table 2. Measured materials temperatures under various outdoor exposure conditions Exposure 45° open back 45° solid back 5° open back 5° solid back Standard black box 5° Heated black box 5° Alnbient air temperature

HI,oC

Acrylic black LO,oC

AV,oC

HI,oC

Acrylic \vhite LO,oC

AV,oC

60 74 67 83 76 103 34

49 52 50 46 61 68 26

53 65 57 61 67 81 30

43 49 46 52 53 83

27 37 33 37 34 44

36 42 38 40 44 64

SELECTING SPECIFIC TEST CONDITIONS OR CONFIGURATIONS Once a location (or series of locations) has been chosen for exposure, the next step, selecting the angle of exposure and type of rack fixture, is all important. The following discussion is a description of the most commonly used exposure angles and exposure devices. The absence of any angle or exposure condition from the list does not preclude it from being a viable way to test materials or systems for durability to weathering. The differences between those listed are basic to the question of the choice of the environment under which the materials or systems are to be tested. Each angle change or change in the type of test rack will influence the radiant energy received by the sample, the wet periods it will endure, and the temperatures it will achieve. To illustrate the effect ofexposure angle and test configuration on temperature ofthe test materials, typical temperatures for black and white automotive acrylics for six different types of exposures at Miami, FL are listed in Table 2. Temperatures presented are the highest and the lowest recorded and the average values from eight readings collected over a period of twelve months. Dramatic changes in temperature can occur in a very short time with wind, cloud cover, time of day or rain. Under Glass Exposures: ASTM G24 under glass exposures are used for materials which normally will not be subjected to all elements ofweather while in service. Test specimens are placed 75 mm (3 in) behind 2.5-3 mn1 (0.1 in) single strength window glass, specified as flat-drawn sheet glass which absorbs radiation below about 310 run and increases in transmittance to approximately 90% at 380 nm. When properly maintained, the glass should transmit a minimun1 of 77% ultraviolet in the range specified, 90% illuminate C (average daylight) and 85% total radiation.

Outdoor Weathering Tests

Test Fixtures: Figure 5 is a test building specifically designed for weather testing window systems. The building is air conditioned and heated to simulate in-service conditions. Window systems in Figure 5 are installed at 90° South facing. But such a test building could be altered to provide other orientations of exposure. To reduce costs of expoFigure 5. Test building. sure, parts of systems are often tested. For example, 152 x 305 mm (6 x 12 in) specimens are mounted on anodized aluminum racks which typically have a 64 mm (2.5 in) mask to provide an unexposed area on a test specimen. Specimens may be exposed without the mask when the entire test surface is to be exposed. Test fixtures or racks may be constructed ofany material which will not interfere with the test, and which is suitable for geographical area in which they will be used. The distance above ground for the lowest section of the rack is dependent on the location. However, test specimens should be mounted at a sufficient height to avoid contact with vegetation and to prevent damage which might occur during area maintenance (grass cutting, regraveling, weed control, etc.). The area surrounding the test site should be free of objects likely to shade the specimens during exposure. Ground cover in the immediate vicinity ofthe racks must be representative of the location; gravel for desert areas (to reduce possible abrasion caused by blowing sand), and low-cut grass for most other areas. Roof top exposure set-ups are excluded from ground cover requirements. Standard Exposure Tilt Angles: Even though most exposure tests are conducted facing South or North at 5 or 45° from the horizontal, many racks are adjustable to accommodate those clients specifying station latitude, 90° or variable angle. All of the tilt angles listed below may be used for either direct or under glass exposures at any facing direction and with solid backing, expanded metal backing or without backing. Figure 6 shows typical exposure racks used at an outdoor weathering site. 0° (Horizontal): Primarily used for environmental etch, and roofing materials. Specimens exposed at 0° normally experience longer wet-time with poor drainage and more severe dirt retention.

Weathering of Plastics

5°: Preferable to 0° as it allows for some drainage and dirt wash off during rains yet provides longer total time of wetness than other angles. 5° is used for testing many materials and is used extensively by the automotive industry for testing paints and coatings. Station Latitude: At latitudes above 35°, station Figure 6. Exposure rack. latitudes provide maximum exposure to solar radiation for fixed angle exposures. ISO TC 61 recommends exposure angles of 10° less than latitude to ensure maximum annual solar radiation. In the Miami area, there is no advantage in respect to total irradiation, and significant differences in test results have not been reported. Station latitude is specified for the exposure ofsome materials designed for solar energy conservation applications. 45°: The angle at which the largest variety of materials is tested is 45°. Specimens exposed at 45° will nonnally experience reduced time of wetness, less dirt retention, mildew and, depending on the time of year, lower temperatures than duplicate specimens at 5°. 90°: Vertical exposures are often used for testing residential and commercial construction materials such as siding, window and door profiles, etc. This exposure angle significantly reduces wet time and nonnally lowers the temperatures of specimens being tested. Vertical racks are designed to off-set test specimens to prevent contamination from wash-down from the specimen mounted above. Variable Angle: Seasonal angle changes provide optimum exposure to radiant energy and nonnally result in higher temperatures, and may be applied to test all materials. The ASTM E782 angle change schedule and irradiation data for variable angle, 5°,45° and station latitude Miami, FL and Phoenix, AZ may be found in Table 3. Client Specified Angle: The above tilt angle listings represent those angles most commonly used; however, any fixed angle or variable angle change schedule may be specified by the client to meet special test conditions. In addition, rack fixtures may be turned to face directions other than due North or South.

Outdoor Weathering Tests

Table 3. Variable vs. fixed angle exposures (MJ/m 2) 26° (FL) and 34° (AZ) North latitudes Month

Location

FL AZ FL AZ FL AZ FL AZ FL AZ FL AZ FL AZ FL AZ FL AZ FL AZ FL AZ FL AZ FL AZ

JAN FEB MAR APR MAY JUN

JUL AUG SEP OCT NOV DEC TOTAL

Fixed angle avera~ e 26/34°

5°

460.83 453.09 455.39 496.89 572.33 650.28 567.94 747.89 516.10 751.45 450.57 744.42

Variable (ASTM E782)

45°

511.08 (45°) 583.21 (45°) 492.92 (45°) 610.74 (45°) 572.33 (26°) 650.28 (34°) 619.52(5°) 720.44 (5°) 604.48 (5°) 836.80 (5°) 564.92 (5°) 848.26 (5°) 586.43 (5°) 780.32 (5°) 492.83 (5°) 758.27 (5°)

442.25 329.70 447.23 410.70 579.86 574.84 619.52 720.44 604.48 836.80 564.92 848.26 586.43 780.32

485.68 731.87

511.08 583.21 492.91 610.74 563.96 692.91 522.08 711.74 450.62 698.94 389.95 650.78 427.73 642.54

562.66 747.35 498.73 630.19

492.83 758.27 470.62 731.24

449.95 693.87 455.01 715.38

513.63 498.31 422.46 380.16 401.83 324.51 6244.37 7081.84

509.61 681.36 441.24 653.04

516.56 700.57 466.77 641.66 488.73 595.51 5735.34 7937.84

436.10 535.38 5859.23 7935.58

470.62 (5°) 731.24 (5°) 509.61 681.36 466.77 641.66 488.73 595.51

(45°) (45°) (45°) (45°) (45°) (45°)

6298.62 8351.68

Variable an2le - percenta2e of fixed angles Location

Fixed angle

ASTME782

FL/AZ FL/AZ FL/AZ

5°/5°

103.4/119.3

26°/34°

110.2/106.4

45°/45°

112.6/106.4

Weathering of Plastics

ASTM G90 Accelerated Outdoor Exposures: The Fresnel-reflector test machine, which is a follow-the-sun rack, has a ten-mirror collector array to concentrate sunlight onto an exposure area as described in ASTM G90. (See Figure 7). The axis is oriented in the North/South direction, with the North pole being altitude-adjustable to account for Figure 7. Fresnel-reflector test machine. seasonal variations in solar altitude at zenith. The machine is equipped with a blower to cool the test specimens. The air is directed over and under the samples by an adjustable deflector along one side ofthe exposure area. This limits the increase in surface temperature of most materials to 10 0 e above the maximum surface temperature that would be reached by identically mounted samples exposed to direct sunlight at the same time and location without concentration. Exposures on these devices may be conducted with or without water application. Specimen spray schedules include day only (rain simulation), night only (dew simulation) or day and night applications. The effective exposure area is 130 x 1220 mm (5 x 48 in); therefore, samples should not exceed 130 mm (5 in) in one dimension. Equatorial Tracking: Like the Fresnel-reflector device, the equatorial tracking device follows the sun through the day and has a mast which allows the entire exposure area to be adjusted for seasonal variations in the solar altitude. The increase in radiant energy compared to a 45° exposure is about 35%. The temperature of the samples is on average higher than those mounted in the same way on fixed racks because of the attitude of the sample to the sun, but will not be above the maximum recorded on anyone day. Typically, this unit is run between the hours of7 AM to 6 PM and locked into a fixed position at night or in very heavy winds. In the desert, spray cycles may be added for controlled wet periods to the exposure program.

TIMING THE EXPOSURE TEST Historically, outdoor weathering tests have been timed by four methods: calendar (days, weeks, months or years), ultraviolet sun hours (UYSH), total irradiation (Joules), or to a predetermined change in either a standard reference material or the actual exposure specimen. A high confidence level cannot be established on test results which constantly change. The variability of weather does not lend itself to obtaining repeatable results. "Average

Outdoor Weathering Tests

weather just does not exist within the span of time for test purposes" (Bores-1976), or "Weather cannot be reproduced" (Helmen and Hess-1979) are just two remarks indicating variability. Poor repeatability/reproducibility should be expected because of differences in climatic conditions between geographical areas and seasonal variations at a specific location. Ultraviolet sun hours (UVSH) nlay have served a purpose years ago but, with today's technology, it is an unrealistic unit for timing exposures. Examining the definition of an UVSH help make its shortcomings obvious. An UVSH is any cumulative 60 minutes when the intensity of inconling solar radiation is above 0.823 cal/cn12• Any measurement over 305 mm (12 in) is a foot. Both statements are true, but few would want to dig a ditch one foot deep using the above definition. It could tum out to be a rather deep hole. A similar problem exists with using total irradiation measurements to time exposures. Neither UVSH nor total irradiation take into consideration the quality of sunlight. It is quite possible to monitor nearly the same amount ofUVSH and total irradiation on a day in December and a day in June. However, for the northern hemisphere, the ultraviolet content of June sunlight is at least twice that of December. Summer sunlight may contain well over three times as much ultraviolet as winter sunlight. Whenever possible, it strongly recommended that outdoor exposure programs be timed either by total ultraviolet or by selected wavelength measurements. Either of these methods will reduce the effects of seasonal variations and improve repeatability in test results. The use of a reference material to time the length of an exposure is common. The materials vary fronl industry accepted standards such as AATCC Blue Woollightfastness standards in the textile field to metals of known corrosion rates for the coatings and steel industries. Exposures are conducted for a given period of time under a specific condition until the standard material exhibits a predetennined color change or measurable degree of degradation. Equally common is the use of a known material specific to the company doing the testing. Included in the population ofeach test or return, these materials act as "clock" for the test lot, e.g., the instruction for return may be for a given lot when the standard material has a gloss loss of 10%. In some cases, reference materials are checked to gage the period of exposure when the test is conducted using a separate timing method of calendar days or sunlight measurements.

CONCLUSIONS AND CAUTIONS Clearly, outdoor weathering programs need to be carefully designed if the results are to provide reliable and relevant data. It is also clear that, with outdoor or natural exposures, there are far more options for exposure, acceleration, and correlation studies than conventional laboratory environments offer. If there is a single point to be gained from this paper, it is to pay

Weathering of Plastics

particular note to the variables associated with outdoor weathering. Plan for them when designing a test and review them when the exposure period is complete. The weathering of materials in a given climatic region for a year or two mayor may not provide an accurate picture of the materials' response to climate, the average environmental factors over centuries. The infonnation gained is related to weather, periods of months or years. In an 18 month period in 1989-91, Mian1i rainfall was more than one meter below recorded averages for the last 100 years. This single factor influenced recorded total wet time by as much as 20% which in tum affected chalk development, dirt retention, mildew growth, etc., for all the materials on exposure at that time. Finally, there are two cautions that should be made in direct relation to weathering programs. First, the caution to ensure proper handling of the test specimens. Attention paid to curing times and packaging for shipping will reduce the chance of having your weathering program tum out to be a test for "durability of materials to shipping." Second and equally important are the choices in evaluation techniques that are used to produce the data representing results from weathering. Without an appreciation for the multiple sets of criteria and conditions under which such things as gloss readings, visual ratings, or color units are recorded, the exercise of testing a product for its durability to weathering can be a waste of time and money.

REFERENCES Bauer, D. 1997. Predicting In-Service Weatherability of AutolTIotive Coatings: A New Approach. Journal of Coatings Technology, (69): 85-96. Bores, M.N.M. 1976. Results of Co-operative Exposure Tests in Different Climates (final report to participants), TNO Delft, The Netherlands. Bourke, K. and Davies, H. 1996. Factors Affecting Service Life Predictions of Buildings: A Discussion Paper. Building Research EstablishlTIent, UK. Fischer, R.M. and Ketola, W. 1994. Surface Temperatures of Materials in Exterior Exposures and Artificial Accelerated Tests. Accelerated and Outdoor Testing of Organic Materials, ASTM STP 1202. Helmen, T. and Hess, E. 1979. The Importance of the Suntest Machine for Accelerated Weathering of Paints, Farbe and Lack, (85): 835-841. Jerrlberg, P., Sjostrom, C. and Lacasse, M.A. 1997. State of the Art Report; TCI40-TSL: Prediction of Service Life of Building Materials and Components. Materials and Structures, RILEM, France. Koch, W. 1980. 15 Years Experience in Outdoor Weathering of Facade Finishes in Florida and Switzerland, ECCA Congress, Brussels. Martin, 1.W., Saunders, S.C., Floyd, F.L. and Wineburg, J.P. 1994. Methodologies for Predicting the Service Life of Coating Systems. NIST Building Science Series 172. Gaithersburg, MD. Masters, L.W. 1986. Prediction of Service Life of Building Materials and Components. Materials and Structures. RILEM, France. Masters, L.W. and Wolfe, W.C. 1974. The Use of Weather and Climatological Data in Evaluating the Durability of Building Components and Materials. NBS Technical Note 838. The National Institute of Standards and Technology, Gaithersburg, MD 20899. Misev, L. et al. 1999. Weather Stabilization and Pigmentation ofUV-Curable Powder Coatings. Journal o/Coatings Technology (71).

Outdoor Weathering Tests

Nichols, M.E. and Darr, C.A. 1998. Effect of Weathering on the Stress Distribution and Mechanical Perfonnance ofAuton10tive Paint Systems. Journal of Coatings Technology (70). Severon, B. 1998. Bring Me Sunshine. Testing Technology International (1). Shirayama, K., Editor. 1993. Principal Guide for Service Life Planning of Buildings. Architectural Institute of Japan. Tokyo. Sjostrom, C., Editor. 1990. Feedback froln Practice of Durability Data: Inspection of Buildings. CIB Report 127. Rotterdam. Sjostrom, C. and Brandt, E. 1991. Collection of In-Service Performance Data: State of the Art and Approach by CIB W80IRILEM 100-TSL. Materials and Structures. RILEM, France. Wicks, Z. W., Jr. et al. 1999. Exterior Durability. Educational Series, Journal ofCoatings Technology (71). Wootton, A.B. 1996. Accelerated Weathering Specifications used in the Polymer Industry. Proceedings ofPolymer Testing '96. Rapra Technology Limited, UK.

A Comparison of New and Established Accelerated Weathering Devices in Aging Studies of Polymeric Materials at Elevated Irradiance and Temperature

Jorg Boxhammer and Kurt P. Scott Atlas Material Testing Solutions, Ger111any and USA

INTRODUCTION The main objectives for testing under simulated environmental conditions in laboratory instrumentation are to conduct the tests under more controlled and accelerated conditions as compared to outdoors exposure. The reproduction ofapplications effects, on the one hand, as well as the precision and the speed on the other, are the key factors of a good accelerated weathering test. A great deal of tin1e and money are continuously spent by industry in scrutiny of these points in an effort to improve the quality of test design. Concurrently, there has been an accompanying strong effort by instrument manufacturers concerning improved and evolving equipment technology. Compared to other simulation techniques, laboratory weathering devices that utilize filtered xenon arc light sources have the advantage of a full spectrum including all wavelengths that exist in sunlight; which, ifproperly filtered, can be modified to provide a spectnlm which closely resembles sunlight. Improved measuring, controlling and calibration techniques ensure that critical physical paratneters are maintained at specific levels. But nevertheless, various types of instrumentation which employ different techniques to control and measure the test parameters in addition to their calibration regimes, must be investigated for their impact on test results. This is especially valid for test methods that are designed to simulate extreme environmental conditions, as for example, for the purpose of predicting the performance of automotive materials for interior applications. Previous experimentation and reports in literature on the lightfastness of automotive textile materials 1 indicate that reliable test results can be obtained for these materials in different types of xenon arc instruments. In the present

Weathering of Plastics

study several widely used automotive polymers were investigated in instruments which employ different technologies to control critical test parameters and correlation of test results will be discussed. Especially in the field of test methods for automotive n1aterials, a major concern is the length of test time necessary for qualification tests. The ultraviolet portion of sunlight is known to have the greatest degradation effect on materials and can be used to accelerate the processes in one of two ways: by increasing the intensities of the wavelengths found in sunlight, or by including shorter wavelengths than those found in sunlight. The latter also increases the likelihood ofproducing results which may never occur during the service life of a real automobile. Specific SAE-standards are based on the latter method. However, experimental work is ongoing for replacement by the methods based on filters that produce a better match for sunlight. 2- 7 These methods are already included in the JASO-standard M346: 1993 - Light Exposure Test Method by Xenon-arc Lamp for Automotive Interior Parts 8 and the revised DIN-standard 75 202: Materials for Automotive Interior Application; Lightfastness Testing and Aging Behavior to Light at High Tenlperatures: Xenon-arc Radiation (method 2: Determination of colorfastness)9 - for automotive interior materials which are based on "sunlight behind window glass" and increased levels of irradiance have already been included. The specified testing conditions listed in Table 1 are similar for the range of irradiance as well as temperature, taking into account the well known systematic differences between black standard and black panel temperature. 10 Previous experimentation reported in the literature on the lightfastness of automotive textiles according to the DIN-test method at varied irradiances 1 indicate that reliable test results can be obtained, provided the sample surface temperature of any exposed specimen is kept constant. But nonetheless, further systematic investigations are still needed and have been conducted on a number of industrial polymeric materials for automotive interior applications in the instruments Xenotest Alpha and Xenotest Beta. The test results will be discussed. The irradiance on specimen area in laboratory instrumentation can be raised by either increasing the power of the lamp or decreasing the distance between the lamp and the sample surface. Both approaches are used in different types of available instruments and both increase the irradiance of all wavelength proportionally. Thus the increased spectrum also contains increased levels of visible and infrared radiation as well as UV radiation. The increased infrared radiation can be problematic. However, over the last few years, instrumentation has been introduced that make it possible to control both black panel (or black standard) temperature, as is traditionally the case, as well as chamber air temperature. This is an important innovation as it allows san1ples ofvarious colors, from dark to light, to be tested in accelerated weathering devices in a similar manner to that of outdoor testing. This feature is accomplished by a combination of the use of sophisticated thin-film coated filters

Elevated Irradiance and Temperature

Table 1. Contents of DIN 75 202, method A (revised version) (JASO M 346 • 93 given for comparison) Specification Spectral distribution of radiation general

DIN 75 202 - method A

JASO M 346 - 93

Sunlight behind window glass (CIE No. 85, table 4 + 4 mn1 window glass)

detailed below 320 nm, %2

Spectral function up to 800 nln <1.1 (from spectral function)

300-400 nm, W m- 2 420 nm, W m- 2 nln- 1

44-62 (up to 150)3 1.0-1.4 (up to 3.4)3

1.5

Irradiance

Black standard temperature, °C Black panel temperature, DC Chamber telnperature, DC

65±3

Relative humidity, 0/0

20±10

Radiant exposure 300-400 nm, MJ m- 2 420 nm, KJ In- 2 nln- 1 Standard reference material

11400-141004 260-320 4 blue wool reference types 5-8

48-162

100±3

89±3 50±5 see applicable specifications

tRevised version (also basis of Gennan proposal for revision of ISO 105-B06) 2E(300-400 nm) = 100% 3High irradiance only for detennination of colorfastness (procedure 2) 40ne test period according to endpoint definition; number of test periods see applicable specifications

which either reflect or absorb excessive infrared light, partially in combination with a new xenon-lamp technology with a reduced IR-portion ofradiation (instrument Xenotest Alpha) and a variable speed blower that modifies the air flow over the specimen surface as necessary. The UV-irradiance range that may be obtained at given black standard and air temperatures can be estimated from apparatus' operating ranges shown in Figure 1 for the Xenotest Alpha and Beta instruments. These operating ranges clearly show the reduced temperature level in the Xenotest Alpha compared to the Xenotest Beta (same filter system) as well as the extended UV-irradiance range. For conducting tests in the Xenotest Alpha at low UV-irradiance according to the temperature requirements in DIN 75 202 (BST - CT = 35K) the filter system has even to be replaced by such with a high infrared transmission (window glass). High sample surface temperatures used in the test methods for automotive interior materials are necessary for simulating the in-service conditions and do increase the acceleration due to enhanced photoinduced reactions. Material specific transition temperatures and thermal degradation processes may affect the reactions especially at those high temperatures. The

Weathering of Plastics

70·

60 50

I-

IIII

lXl

20 10

A - minimum air speed B - maximum air speed

W-Irradiance (Wlm')

Figure I. "Operating ranges" in Xenotest equipments Alpha and Beta - range ofirradiance at unchanged temperature conditions.

increased criticality oftemperature could cause slightly changed surface temperatures to produce significantly different ageing effects in the exposed materials. II Tests at varied temperatures may provide valuable information on those phenomena and are therefore included in the present study, Thus the following work is intended as a three-step experiment to provide data from comparative tests on industrial polymeric materials for automotive interior applications in different kind of established instrumentation. The tests are done at the "normal irradiance level" and in new types of instruments, also at high UV-irradiance as well. In addition, the effect of temperature is investigated at three different temperatures. All tests conducted are based on the German specification "Interior materials in motor vehicles - Determination of lightfastness and aging behavior to light at high temperatures: Xenon-arc radiation".9 The color change is the performance property examined which is instrumentally measured.

Elevated Irradiance and Temperature

Table 2. Specimen identification Sample code PUl PU2

Description (as submitted for the study)

Color

Polyurethane

gray

Polyurethane

gray black

Remark fabric behind woven fabric behind

PU3

Polyurethane

PVC

Polyvinylchloride

light gray

PPl

Polypropylene (stabilized)

tan

foam behind - 2 mm

PP2

Polypropylene (unstabilized)

foam behind - 2 mm

PP3

Polypropylene (unstabilized)

black light gray

woven fabric behind

EXPERIMENTAL DESIGN MATERIALS

Seven industrial materials for automotive interior application were submitted by an automotive supplier. Table 2 shows the material identification, each with a generic description code which is used throughout the text. The polymer samples were mounted on white cardboard and exposed with fleece backing in accordance with the requirements of DIN 75 202. The samples were then placed and exposed in the specimen holder which is considered standard for each of the different kind of test instrument. TEST INSTRUMENTS

All tests were conducted in different types of xenon-arc equipment that are designed to meet the requirements specified in DIN 75 202. The specific types of equipnlent are listed in Table 3. The Alpha and Beta instruments are equipped with a measuring device (XENOSENSIV) which has its sensors in the sample plane for measuring and controlling the UV-irradiance (300 - 400 nm) and the black standard temperature. These instruments have also been used for the tests at varied in adiance and temperature. The Ci-series Weather-Ometers are equipped with a narrow band light monitoring and controlling system; at 420 nm for this particular test method. The controlling black standard thermometer is mounted so that its sensing black surface is in specinlen plane. 4

TEST METHOD AND CONDITIONS

All tests were based on DIN 75 202. The specified conditions are shown in Table 1. The various test conditions are listed in Table 3. The instrument comparison tests (1 to 6) were all conducted at a low UV-irradiance level (except for test 6 - slightly increased irradiance level).

Weathering of Plastics

Table 3. Equipment and testing conditions Irradiance Equipment

Filter system

Xen Alpha HE Xen Beta LM

Suprax + wg Suprax + X. 320

Xen Beta LM Ci3000WOM

Suprax + X. 320 Boro S/Sodalime

Ci3000WOM Ci4000WOM

Quartz/Bo S/wg I

Xen Xen Xen Xen Xen Xen Xen

Suprax + wg Suprax + X. 320 Suprax + X. 320 Suprax + X. 320 Suprax + X. 320 Suprax + X. 320 Suprax + X. 320

Alpha HE Alpha HE Beta LM Alpha HE Alpha HE Beta LM Beta LM

Temperature, ·C

0/0

48 48

65 65

100 100

20 20

l.l

48 (48)

65 65

100 100

20 20

3 4

1.1 1.4

(48) (61 )

65 65

100 100

20 20

5 6

96 144 96 144 144 96 96

65 65 65

100 100 100 90 110 90 110

20 20 20 20 20 20 20

55 75 55 75

I:: Vi

0()'o2m~ -,

280

300

•

~\ t ~

;EI rf,J

.:7 . :120

7 8 9 10 II 12 13

'---"1\

S.:Sodnlinc (Ci3CXXJ)

""""""d]oo:~oo"m

I 2

- - - - Supmx+ X:l1ochrom: 320 (Alphai&ln)

- • - • - Boll')

Test no.

BST

- - Supra:..;' \\iudo\\" glass (Alpha)

1.0

Lab.

Chamber

1,4 - .....

~ o

300-400 nm W/m 2

420nm W/m 2nm

340

I 360

\:J

380

400

WavelengJJl (11m) Figure 2. Filter system (equipment) and spectral irradiance in UV-range of radiation.

The tests for studying the effect of varying irradiance were carried out at increased irradiance by factor 2 (Xenotest Alpha and Beta) and by factor 3 (Xenotest Alpha). The tests for study-

Elevated Irradiance and Temperature

ing the effect of varying temperature were also conducted in the two types ofinstrument operated at different levels of irradiance and at an increased or reduced temperature level (BST and CT by ±10K compared to the conditions specified in DIN 75 202). All filter systems used in the different type of instrun1ents and tests meet the requirements specified as a spectral function in DIN 75 202. The graph in Figure 2 shows a comparison of the spectral power distribution (SPD) of the Xenotest and Atlas Ci-series instruments when each are operated under normal irradiance and respectively controlled as described above. TEST PERFORMANCE The exposure tests as well as the evaluation of test results were conducted in three laboratories (Table 3). MATERIAL PROPERTY The change in appearance n1easured as change in color was the parameter of interest. Color change in the specimen after light exposure was determined instrumentally using a spectrocolorimeter, illuminent D65, 10° observer, in the CIELAB color space. Three measurements per specimen were averaged prior to calculation of color. Color differences were based on the difference between the exposed surface of any specin1en and its corresponding value n1easured prior to exposure. Color measurements were carried out in increments as described below. EVALUATION INCREMENTS All tests were conducted on the basis of"light periods" as specified in DIN 75 202. This "endpoint" defined by an instrumentally determined specific color change of i1E* == 4.3±O.4 (CIELAB, D65, 10° observer) on the European Blue Wool reference standard type 6. The radiant exposure needed to reach this contrast on the reference was established for each equipment at "normal conditions" of irradiance and temperature prior to the test. The tests were then conducted in increments up to a maximum radiant exposure equal to six light periods. For each light period a new reference was exposed adjacent to the specimens for the purpose of determining whether the apparatus was operated within the desired range. The tests at increased and reduced temperatures were conducted for the same radiant exposure increments established at "normal" temperature. EVALUATION OF TEST RESULTS In this study, statistical correlation of the test results in different types ofequipment as well as at varied conditions is examined. Correlation, as applied to a weathering test, refers to the agreement of test results from different instrulnents and/or varied conditions in rating perfor-

Weathering of Plastics

mance differences of various samples of inherently different durability. The experimental devices and varied conditions were evaluated for agreement using Pearson's correlation coefficient, r. The Pearson correlation coefficient is a statistic that provides a normalized and scale free measuren1ent of the linear association between two variables. The higher the correlation coefficient (maximum == 1), the stronger the association between the two variables.

RESULTS AND DISCUSSION TEST COMPARISONS If two instruments were perfectly correlated, the calculated Pearson's coefficient of correlation, r, would be == 1, and the color change (~E*) of all the materials tested in each machine, plotted against each other, would describe a perfectly straight line as illustrated in Figure 3 (line A). Therefore, proximity of the calculated correlation coefficient to unity (1) and the plotted data to a perfect straight line are both sound and visual means to con1pare independent test results.

INSTRUMENT COMPARISON Test results from all investigated materials were used to compare the different kind of instruments. Figure 3 is a plot of the results from two of the instruments: Ci3000 and Ci4000 Weather-Ometers and for two test increments. In both cases the cOITelation coefficient is near to I' == 1. Table 4 shows the correlation coefficients of results from all the instruments for the same test at the longest test increment. This is representative of similar analyses which were done on the results of all materials at each test interval. All summary tables were not included because of space considerations.

Table 4. Comparison of equipment • Pearson's correlation coefficients for ilE* and test increment 70 MJ/m

Alpha - Test I Beta - Test 2 Beta - Test 3 Ci3000 - Test 4 Ci4000 - Test 6

Alpha Test 1 1 0.919 0.793 0.819 0.846

Beta Test 2 0.919 1 0.907 0.899 0.895

Beta Test 3 0.793 0.907 1 0.979 0.969

Ci3000 Test 4 0.819 0.899 0.979 1 0.962

Ci4000 Test 6 0.846 0.895 0.969 0.962 1

Analysis of results from repeated tests in Xenotest Alpha and Beta are shown in Table 5. Again, the high I' values are a good indication that test repeatability can be accomplished.

Elevated Irradiance and Temperature

Table 5. Repeat tests - Xenotest Alpha and Beta; Pearson's correlation coefficients calculated for ~E* and test increment 29 MJ/m 2 Alpha Test 811

Alpha Test 8/2

Alpha - Test 8/1

0.899

0.983

0.968

Alpha - Test 8/2

0.899 0.983 0.968

0.956

0.969

0.956 0.969

0.993

Beta - Test 9/1 Beta - Test 9/2

4,5

I • 620 KJ/I1-r

4 ............

3,5

r=l U

.....

._--,-~

i/ ..• "';;iii";'-'"

__.980 K.J/n-r!

2,5

Beta Test 9/2

-- ~'l' ;;;+-/' ; ..!......

..' l

.........

y = 1.0335x + 0.1379

3 0 0 0

Beta Test 9/1

=0.975

-r '

2·

................. -

1,5 .

'--__ .oo_,t, __ ..

;~..--_. .

..7 ·

··1·, ....

_.._.._ ~

y = UZ93x + 0.0381 r = 0.99

0,5

0 0

0,5

1,5

2,5

3,5

4,5

Ci4000 Figure 3. Comparison of test results - Ci3000/Ci4000 (tests 5 and 6); determination of Pearson's correlation coefficient for t1E* and two test increments.

EFFECT OF VARYING IRRADIANCE There is considerable interest in shortening the test time of materials that are considered for production. One of the most obvious ways to do so is to test at higher irradiance levels. Naturally, this option has value only when the test results are not significantly altered from those tested at "normal" or traditional irradiance levels. Tests were conducted at high, normal, and low irradiance to investigate the impact of varying irradiance. The results are summarized in Figure 4 and Table 6. Figure 4 shows ~E* values averaged for all materials at different levels of irradiance and three test increments. Table 6 contains the correlation coefficients for two

Weathering of Plastics

2.5

Alpha 48 W m-

Alpha 96 W m-

Alpha 144 W m-

Beta 48 W m-

Beta 96 W m-

1.5 -iC

W <J

1 0.5

13.5

H, MJ m-2 Figure 4. Influence of irradiance on correlation for three test increments;

~E*

- mean values for all materials.

Table 6. Influence of varied irradiance on correlation. Pearson's correlation coefficient for ~E* and two test increments Irradiance, W m- 2

144

1\£ *; test incren1ent 27 MJ m-

Alpha - Test 1

Alpha Test 7

Alpha Test 8

Alpha Test 2

Beta Test 9

0.927

0.901

0.886

0.870

0.960 1

0.974 0.991

0.976

1 0.997

0.997 1

0.917

0.919

0.935

0.803 1 0.983 0.942

0.794 0.983 1 0.939

0.921 0.942 0.939 1

0.927

Alpha - Test 8

0.901

0.960

Beta - Test 2

0.886 0.870

0.974 0.976

Alpha - Test I

Alpha - Test 7

0.928 0.917 0.919 0.935

Alpha - Test 8 Beta - Test 2 Beta - Test 9

Alpha Test 1 Alpha - Test 7

Beta - Test 9

48 2

0.991 0.983 ~E*; test increment 92 MJ m- 2 0.928 1 0.803 0.794 0.921

0.983

Elevated Irradiance and Temperature

9,(Xl

....

8,00 ". 7.00 w

£:.;

6.00 5.0)-

4,00

3.lXl -

2,00 I,CX) .

0,00

,,-_.-,-_.~. --'--~--,-_.~

14,2

28) H [MJ/m'}

56,7

85,S

Figure 5. Color change, D.E*, as a function of radiant exposure at varied temperatures - examples.

test increments. The results indicate no significant adverse effect, implying that this option may well be viable. EFFECT OF VARYING TEMPERATURE

Unlike the experiment of varying irradiance, the question here was not whether this could be employed as a means of shortening test times. Rather, this experiment was expected to show the opposite, as it is well known that materials' sensitivity to heat is widely variable. Since any difference in temperature would have unequal impact on specimens, elevating temperatures to shorten test times is not viable. Figure 5 - ~E* vs. radiant exposure - substantiates this argument by showing a significant impact on the degradation rate especially for the material PUI, while having little effect on the rates of the PVC and PPI materials. For the PPI and PVC materials which are presumed to be relatively insensitive to heat, all ~E* values from tests at the various temperatures are within the limiting curves. MATERIAL STABILITY

The investigated polymeric materials for automotive interior application cover a wide range of stability as indicated by color change under the applied severe exposure conditions. Figure 6 shows the total color difference for all materials as a function of radiant exposure from tests in the four instruments. Highest and lowest degradation rates are valid for the polyurethane

Weathering of Plastics

•

I'lil

... • .....

• 'Pv'C

.. Pf'l

•

PP!

•

ItJJ

," ""'111

,.' ':'

,..:

. •

-...'

~I'-. I

HIMJIlll'j

Alpha - to:.t 1 Beta· test 2

'.... .., ~o

• lHl

100

H IMJ/""j

(;

';=3

If (\-LJ!m'l

Ci3000 - test 5 Ci4000 - test 6

Figure 6. Polymeric materials - color change i1E* as a function of radiant exposure.

system PUl, the stabilized polypropylene system PPl with all other material systems ranking between. The different degradation rates may not only be based on the polymer substrates themselves but may be influenced by different surface temperatures due to the color and associated infrared absorbing quality as well as "system design" of the material. Considering each instrument individually, and examining the materials' observed aging characteristics, it can be seen that the extremely disparate aging characteristics of the polyurethane material PU I and the stabilized polypropylene PP I are reproducible within acceptable limit. But significant differences are indicated for materials that performed in the intermediate range.

CONCLUSIONS Results from the high irradiance tests were promising, suggesting that there is little risk associated with its use which may offer industry means to shorten test times. However, while the test samples were varied, the population was still limited and may not be universally representative. Also, it is important to keep the other parameters, especially temperature, unchanged, while increasing irradiance.

Elevated Irradiance and Temperature

Temperature sensitivity of materials underscore the need for close control of both black standard/panel and chatnber temperatures. The test results suggest that temperature variations will alter the relative rates ofdegradation depending on the materials' sensitivity to heat. For a few materials, the coefficient of correlation between Xenotest and Ci-series models, though acceptably high for weathering tests, was somewhat lower than the general population. It is believed that these materials are particularly sensitive to the small but potentially important differences in SPDs shown in Figure 2. Further investigation is necessary to prove if this is so, and ifit is; a continued effort to reconcile between model variations is justified. Othetwise, there was generally good correlation among the test results from the evaluated weathering instruments, which included those using different technologies to measure and control critical test parameters

REFERENCES 2 3 4

8 9 10 11

J. W. Stuck, Experiences with, and Ne\v Options Provided by, High-speed Exposure, Symposium "Colorfastness testing in textile industry", DEKJAtlas-Xenotest, Gelnhausen (Gemlany); 22.10.1996. B.M. Reagan, Accelerated Lightfastness Testing of Disperse Dyes on Automotive Fabrics, Midwest Section's ITPC Committee; 1993 Intersectional Technical Paper Competitions, Dec. 1993; Vol. 25, No. 12 (1993), p. 25 to 32. L. A. Bard, USA Automotive Interior Material Testing; Yesterday, Today and Ton10rro\v; 1993 IPC-7 SAE Conference for Pacific Rim Countries, Phoenix, AZ, Nov. 1993. Y. Watanabe and 1. Matsuoka, Study for Good Correlation and Acceleration on the Artificial Fading Test Methods for Automotive Interior Fabrics; International Synlposium on Auton10tive Test Procedures for Interior Trim Materials; Willian1sburg Virginia, August 17 - 18, 1989. L. Crewdson and K. Scott, A Comparison of Experimental High Irradiance and Standard SAE Weathering Tests for Automotive Exterior Materials; SAE Technical Paper Series 940855; SAE International Congress & Exposition, Detroit, Michigan, March 1994. L. Crewdson, Correlation of Outdoor and Laboratory Accelerated Weathering Tests at Currently Used and Higher Irradiances - Part II, 1st International Symposium on Weatherability (ISW); Materials Life Society, Japan, May 12, 13 1992. L. Crewdson, Correlation of Outdoor and Laboratory Accelerated Weathering Tests at Currently Used and Higher Irradiance Levels - Part III; 2nd International Symposium on Weatherability (2nd ISW); Materials Life Society, Japan, Sept 27 - 29, 1994. JASO M346 Light Exposure Test Method by Xenon-arc Lamp for Automotive Interior Parts; 1993. DIN 75 202 Materials for Automotive Interior Application; Lightfastness Testing and Ageing Behavior to Light at High Temperatures: Xenon-arc radiation. 1.Boxhammer, D. Kockott, and P.Trubiroha, Black Standard Thermometer - Temperature Measurement of Polymer Surfaces During Weathering tests, Materialprt'ifialg, 35 (1993) 5, p. 143 to 147. P. Trubiroha, Alterung polYlnerer Werkstoffe beim Bestrahlen und Bewittern; Industrie-Lackierbetrieb, 58 (1990), 5, 180 - 184.

Current Status of Light and Weather Fastness Standards. New Equipment Technologies, Operating Procedures and Application of Standard Reference Materials

Jorg Boxhammer Atlas Material Testing Solutions, Gern1any

SYNOPSIS The effects of weather on the color and other properties of materials are of considerable technical and commercial importance. There is a need to gain information by accelerated procedures. The conditions for the procedure of these tests need to be stipulated carefully and precisely. Based on this, a number of international standards for the testing of plastics, paints/varnishes and textiles have in the past ten years been revised or elaborated and have been published. Objectives, contents and pending questions are being discussed. Apart from these international basic standards, there is quite a number of test specifications which various national organizations of the automotive industry developed for specific automotive applications of technical materials, and which have today gained worldwide impOltance. The specifications already available or in preparation - all based on the latest findings - are being cOInpared here. The aspect of the "detennination of the thermal stress effect on sample level", which is ofmajor in1portance above all for tests at a high ten1perature level, will be discussed in detail. The mentioned procedure developments and precise detem1inations of test conditions were accompanied by further technical developments of the equipment used for these tests. It is pointed out that by presenting the linkage between radiation and temperature in fOlm ofoperating areas, the possibilities and limits of variable parameter settings in certain apparatus systems can be made transparent to the user. Certain standard reference materials (SRM's) were adopted by some standards to check whether the used xenon-arc apparatus is working properly. The application of these SRM's to

Weathering of Plastics

verify the spatial uniformity in equipment as well as to check the effect of selectively altered stress conditions is demonstrated by means of test results obtained in different Xenotest machines.

STATUS OF STANDARDIZATION Some international committees responsible for certain groups of materials have in the last 10 years thoroughly revised a number of basic standards for the testing sector of light exposure and weathering of materials, which have in the meantime been published or released for publication. For the light exposure and weathering tests under simulated conditions in apparatus, this applies in particular to standards for plastics, paints/varnishes and textiles listed in Table 1.

Table 1. Status of international standards ISO/TC

Standard ISO 4892 part 1 part 2 part 3 part 4 ISO 11341

ISO/CD 11507

Title

Status

Plastics - Methods of exposure to laboratory light sources General guidance Xenon-arc sources Fluorescent UV-lalnps Open-flame carbon-arc lamps

established in 1999 published in 1994 published in 1994 published in 1994

Paints and varnishes - Artificial weathering of coatings and under publication exposure of coating to artificial radiation - Exposure to filtered Xenon-arc radiation Paints and Varnishes - Exposure of coatings to artificial published in 1997 weathering apparatus - Exposure to fluorescent ultra-violet and condensation conditions

ISO 105

Textiles - Tests for color fastness

part B02

Color fastness to artificial light: Xenon-arc fading lamp test published in 1994 Color fatness to artificial light at high temperatures: Xenon-arc published in 1998 fading lamp test

part B06

The essential objectives of standardization can be summarized as follows: Formal objectives 1.1 Reduction of the variety of standards by con1bining the test methods for light exposure and weathering (realized in the standards for plastics and paints).

New Equipment Technologies

1.2

Assignment ofunifonn test conditions for all testing systems in the market in order to generally improve the comparability oftesting conditions (see standards for different radiation systems), 2 Content-related objectives 2.1 Improvement ofthe correlation oftest results in apparatus with the material behavior in practical application by simulating natural stress conditions more accurately (with the acceptance of reduced time acceleration). Note: Basic standards apply to any kind of material and material properties. 2.2 Improvement ofthe comparability and reproducibility oftest results in apparatus by more accurate determination of test conditions and of the relevant Ineasuring systems. Despite an overall different respective unsatisfactory realization of these objectives for the specifications listed in Table 1 (this refers in particular to the textile standards), considerable progress has been achieved in particular for the "testing with filtered xenon radiation". The fact that this radiation system is most suitable to simulate the spectral distribution of solar radiation in the UV and visible spectral range, is nowadays uncontested worldwide. The most important stipulations in this new standard generation refer to the parameters "radiation" and "temperature" and include the following details: 1 Definite perfonnance target ofthe spectral distribution ofthe simulated radiation as the relative radiation function in the wavelength range up to 800 nm and differentiation in the UV range up to 400 nm according to the indications about the spectral energy distribution of global radiation in CIE no. 85 (1989).1 2 Detennination of the irradiance level up to 800 run (for the purpose of reference) ofthe UV range up to 400 nm through the relative spectral energy distribution, based on the value for testing purposes recommended in CIE no. 20 (1972).2 3 Detailed description of measuring systen1s (on san1ple level) for the irradiance and for a ten1perature typical of the thennal stress situation. The present methods ofwide (300-400 run) and narrow (e.g. 340 run) band measurement are well accepted, the discussion on temperature measuring is still going on. Based on the same principle - measurement of temperature of a blackened n1etal surface - so-called black panel and/or black standard thermometers are used, which under identical stress conditions indicate systematically different temperatures. 3,4 The probleln of temperature Ineasurement technique is presently worked on by ISO/TC 61 within the framework of a Technical Report. As agreed, the ISO specifications compiled in Table 1 are adopted by the European Standardization Committee CEN as European Standards, which, in tum, become automatically national standards for the member countries of the European Community, i.e., they will replace the presently applicable national specifications.

Weathering of Plastics

Apart from this, there is a worldwide tendency in standardization of the majority national standards organizations to adopt these International Standards directly as national standards or at least to increasingly use them as a basis for national specifications. Besides the international basic standardization for test procedures concerning the light exposure and weathering in equipment, especially intensive efforts of the automobile industry to develop uniform testing methods for specific requirements, have gained outstanding importance among the variety of standardization work for different material applications. This work was and is primarily done in national organizations of the automotive industry in Europe, the US and Japan. Table 2 shows the published specifications resulting from this work and corresponding to today's state of knowledge. Some of the standards are presently already under revision with respect to the experience gained by now from their practical applications.

Table 2. Status of automotive standards Country

Gennany

Organization

Standard Automotive Interior Materials

DIN 75202 - Detennination ofcolor fastness and aging behavior of interior materials in motor vehicles; xenon-arc lamp test DIN/FAKRA DIN 75220 - Aging of automotive components in solar simulation units

Status Standard (1988) in revision approved as Draft Standard 1994 published 1992

USA

IFAI/SAE

SAE J 2212 - Accelerated exposure of automotive interior trim published 1992 components using a controlled irradiance air-cooled xenon-arc apparatus

Japan

JASO

JASO 1\1 346-93 -light exposure test method by xenon-arc lamp published 1993 for automotive interior parts

SAE

SAE J 2019 - Accelerated exposure of automotive exterior published 1994 materials using a controlled irradiance air-cooled xenon-arc apparatus

Automotive Exterior Materials USA

The German standard DIN 75 202 for material testing was supplemented in the last few years by DIN 75 220, which refers in particular to the testing ofcomponents and complete vehicles. The aim is to determine the influence of construction and respective specific stress situation on the deterioration. The testing conditions are specified for both interior components as well as for car exteriors. The large radiation areas needed for these tests are presently realized for economical reasons with special metal halide lamps. The radiation simulation is

New Equipment Technologies

also based here on the spectral energy distribution of the CIE no. 85 and of the global irradiance according to CIE no. 20. Defined indoor/outdoor as well as day/night conditions can also be corrlbined with test cycles. The development of testing procedures for the specific requiren1ents of the multitude of technical materials in car interiors and exteriors is in the US and Japan linked to the general shift from carbon arc to filtered xenon-arc radiation. The listing ofspecifications clearly shows that the development oftesting procedures especially for the extreme thermal stress conditions inside a car were to the fore ofall activities. The practical application ofthe first specifications for the testing of car interior trim materials as well as now also for car exteriors have provided empirical values which have already been entered into the more recent standards. With a generally increased temperature level for the testing ofcar interior trim materials, the possible measures to enhance time acceleration by simulated radiation the following possible measures were and are controversially discussed especially in the US and Japan: • application of shorter wavelength radiation than existing in global radiation • considerably increased level of irradiance at an application-oriented adaptation of the edge limitation of radiation. Today's experience has already confirmed the risk of insufficient con elation when applying the first mentioned measure S,6 and was allowed for in the SAE J2212 by adapting the spectral distribution to the edge limitation of global radiation at a limited increase of the UV irradiance for compensation of the testing time extension. The direct adaptation of the simulated radiation to the spectral distribution behind window glass as an extreme stress situation inside a closed car was in Japan taken as a basis for the process development with respect to a good correlation. For the achievement of the wanted time acceleration, they use a strongly increased level of the UV irradiance. 7 The test results available today for both, the materials used inside and outside ofautomobiles, lack uniformity 7-9 and do not allow any definite statement concerning the admissibility of restrictions of such test conditions. Further systematic investigations are needed here. The explicit determination of relative spectral energy distribution in accordance with CIE no. 85 as a basis of simulated radiation was - in accordance with ISO standards - entered into SAE J 2212 and DIN 75 220 as well as in connection with the DIN 75 202 revision. In the latter, the term of "spectral energy distribution behind window glass" was clearly defined, and the visible range of the spectrum was included in the spectral function (see Table 3). A comparative sumtnary of the most important features of the recent standards for the testing of automotive interior materials is contained in Table 4. Apa11 from the spectral energy distribution of radiation and the level of the UV irradiance, there is correlation on one hand and divergences on the other, also for the parameters of temperature and hun1idity. The lack of uniformity in temperature n1easurement and 4

Weathering of Plastics

Table 3: Relative Spectrallrradiance for Daylight Behind Window Glass (DIN 75 202-Revision; Basis CIE No. 85 (1989)/4 mm European window glass) Wavelen2th, nm <290 <300 280-320 320-360 360-400 400-520 520-640 640-800 <800

Relative spectral irradiance, l 0 <0.05 <0.1 3.0±0.85 5.7+2.0/-1.3 32.2+3.0/-5.0 30.0±3.0 29.1±6.0 100

0/0

IThe spectral irradiance between 290 and 800 nm is defined as 100 %

indication mentioned for the ISO standards also applies here. Especially the high temperatures applied as well as the thermal stress level may affect the test results to a decisive extent. The realization of identical surface temperatures of any exposed materials is actually of major importance considering the prime objective of correlation as well as comparability and reproducibility of testing results. The situation and the prerequisites for an approximate implementation of this requirenlent in equipment technology are shortly outlined hereafter. Under the influence of solar radiation on surfaces, each specimen is expected to achieve a different surface temperature during the exposure, depending on its absorbing and insulating properties. This thermal differentiation between specimens is caused by the total radiation respectively mainly influenced by the radiant infrared portion of the sunlight spectrum. Only by using an artificial light source containing infrared radiation as given by the continuous xenon spectrum (properly filtered) can this differentiation be simulated in the laboratory. The measurement of individual surface temperatures is not practicable in equipment testing. The generally applied measuring method for the characterization of the thermal level was already referred to before. The construction differences and measuring experiences with the Black Panel Thermon1eter (BPT) and the Black Standard Thermometer (BST) are dealt with in literature. 3,4 As a matter of fact, the BST does not merely provide a reference point for temperature control and test nlethod comparison - for this, a uniform BPT would be sufficient - but the indicated temperature (leaving aside exceptions 3) characterizes the maximunl surface temperature of a thicker polymer sample with poor thermal conductivity.

New Equipment Technologies

Table 4. Contents of automotive standards Specification

DIN 75 202 (revision, condition A)

SAE J 2212

JASO M 346-93

Spectral distribution of radiation general details

sunlight behind window glass sunlight outdoors (basis CIE sunlight behind window glass (basis CIE no. 85) no. 85) Spectral function up to 800 nm Spectral function up to 400 nm _

below 320 om, % I

< 1.1 (from spectral function)

5A± 1.8

light only

light/dark 228/60 min

45-60 1.0-1.4

80±1 97±3 62±2/38±2 50±5/9515

Phases Irradiance, W m- 2 300-400 nm 420 nnl Black standard temp., °C Black panel temp., °C Chamber temp., °C Relative humidity, % Radiant exposure 300-400 nm, MJ m- 2 420 nm, KJ m- 2 Standard reference materials

20±10

1.5

48-162

89±3 50±5

12-162 250-300

see applicable see applicable specifications specifications Blue wool references, Types 5 Blue wool references, Type L2 and L4, Polystyrene to 8

IE (300-400 nm) = 100% 2Method 3: Definition of one "light period"; Blue wool reference, Type 6 - ~E* (CIELAB 065/10°) Number of light periods see applicable specifications

4.3±0.4

With an additional explicit definition and adjustment of a lowest surface temperature of a sample, the surface of which is directly absorbing and reflecting, also the intermediate temperature of any surface can - at unaltered limiting values - be considered as constant (as a matter of fact, both the spectral distribution of radiation as well as the spectral characterization of the individual surfaces have to be allowed for). The white standard thermometer (WST) which ISO 4892-1 and ISO 11341 specify for the measurement of this lowest value - without defining any temperature - and which from its design is identical with the BST has so far not been implemented in practice, whereas the chamber temperature (CT) which is already measured in equipment and which is better suited for regulating processes, is used as theoretically lowest surface temperature.

Weathering of Plastics

The basic standards do not yet specify this additional temperature value which closely restricts the temperature range of any surface. This requirement has, however, already been entered in the DIN 75 202 and in the SAE specifications (see Table 4). In connection with the utilization of Standard Reference Materials (SRM's) for the light exposure and weathering testing in equipment no progress has been achieved in the last 10 years of international discussions. This discussion does not refer to the blue wool references which have long since been applied in the light fastness tests of textiles. Regardless of the international situation, certain blue wool references and a plastic standard (clear polystyrene) to verify whether the used xenon-arc apparatus is working properly, were included in the SAE standards. The blue wool reference standard type 6 is used in DIN 75 202 for the detennination of a "light period" (colorimetrical detem1ination L\E * (CIELAB

D65/100) == 4.3±O.4").

STANDARD REQUIREMENTS AND EQUIPMENT TECHNIQUE The contents ofthe standards under discussion reflects practiced testing techniques in apparatus. The different approaches in process development and the n10re precise determination of parameters to in1prove the con1parability and reproducibility of results have considerable technical consequences for the equipment manufacturer. Flexibility and precision: based on these requirements, one proceeded to new and advanced developments in the field of lamps and filters as well as in the area of sensor technology and data communication and documentation. The results have been consistently realized in today's testing systems. The essential requirements realized in equipment technique refer to the factors ofradiation and temperature. The technologies used are shown in Table 5. Their application in a flexible range of systems implies • the measurelnent (and control) of the UV irradiance (300 to 400 nm) and of the Black Standard Temperature (BST) at a close spatial assigrunent on the level of the exposed specimens; • largely constant spectral irradiance through XENOCHROME systems which due to their dielectric coatings have a reflecting effect (this was already reported); 10 the spectral irradiance for the wavelength range up to 800 run is shown in Figure 1 (those ofeIE no. 85 for comparison); • widely adjustable UV irradiance up to the high energy range with a comparatively largely reduced heat effect through "cool xenon radiation"; result of the development of a new so-called OMEGA xenon lamp. In standardization, the requirements for the radiation and temperature parameters are indicated with respect to their effect. According to the portion of radiation which is decisive for

New Equipment Technologies

1(1

------" XerodlrCNfle 300

•.__.. 01[85

E~ Wm-~nm ............

:w."De"""'~O

. . . .-----..f

......--~-~-----+-----+-----+----200

IIIQ

Figure 1. Xenotest Beta LM - Spectral energy distribution.

photochemical processes, the irradiance is primarily determined for the UV range. As to the surface temperatures which are also decisive for the degradation processes, such level is prescribed that is still considered acceptable and occurs in the specific application.

Weathering of Plastics

l~O

\10

100 90

::::=-::: a

-""'"

"...

.--~

-----

6-0

ccnar(ll4d

..0

:so 70

~dltd

100

WJT

:::::---.~~ ............. ~

:--.::::::--...r::::::- -----.... -

.--.,

~ ~~ ~ -~ ~~~

"-..

-.......

E~W.hll

Figure 2. Xenotest 1200 - temperature vs. UV-irradiance. Equipment: Xenotest 1200; Filter systeln: 3 window glass; Specimen area: stainless steel; Relative humidity: 20 % L - Low air stream H - High air strean1.

In equipment technology the two parameters cannot be considered separately over the total spectrum of the filtered xenon radiation. With a spectral energy distribution unchanged within wide performance limits (setting range of irradiance) the interrelationship between the surface temperature measured under equilibrium conditions and a gradually changing UV irradiance can be described as almost linear. This is shown in Figure 2 by the increase of the limit temperatures BST and WST at constant chamber temperature (Figure 2 upper) by the decrease of the CT necessary for an unchanged BST with an increasing irradiance (Figure 2 lower).

New Equipment Technologies

waveleng1l1 ra'l\g~

...

20.0

2CJo. ':7Cnrn

f·S

5.-1

200- 290n:n 2~O· :lDnm

19.1 38.1 11.7

23.2

44.7

320 - ~6()~rt1

,S.o

8~A

~OOM1

107

11~

1.8-410

300-

~20nm

36Q.

29O.800nm

"* *E

........

:15.0

- ------ -- - .....

~Kf,.1OO0V

Xe 2000

..•

300· WOnm

1.839 1.828

80C • 1C()Onm

3.412

~O· ~OOnTtt

.. ..

3~.3

1.828 ._v.•-----_·

2.~

229

(J:I

of"'!

~O.O

1:1 tu

t..

~~ ~L~~Vu~}~u WVJ j

t. .....

5.00

o ~

.don

700

600 Wav~

! ~ ngt tl

600

rnmJ

Figure 3. Xenon lamps Xe 2000 and NXe 2000U - spectral energy distribution.

Since the ratio between UV irradiance/total ilTadiance is subject to a number of influences, this also applies to the assignment of the UV irradiance and the surface temperature. The modification of the chamber temperahlre which is partly admissible in order to compensate these influences, implies automatically the risk of increasingly different surface temperatures, especially in the case of mostly reflecting surfaces. In today's apparatus technology the assignment of a specific UV irradiance and a specific surface temperature (through the limits ofBST and WST leT) is ensured within a limited setting range by flexibly varying the rates of the air flow which is led over the specimens (Figure 2). For adjustment values modified by "factors" additional measures have to be taken. We already pointed to the "cool xenon radiation" method practiced in the Xenotest Alpha LM High Energy unit. Figure 3 shows the largely unmodified spectral energy distribution up to 800 run and the drop which occurs already in the range up to 1000 run (table values) as compared to a conventional system.

Weathering of Plastics

~. X 300

...

A· X 210 (QJte- <.ytlr$ QJOf2)

(cute- QI1Ir'll3 1pEd(1 w-

.J'~

g(l'»

- C .. 'X31()(QJta'~IrQ;'f'pedoW~ l1a$) .....--

--DlN'~202

100-

----SAEJ ~212

~~ ~

( V(

V V

dT Spa!ilC

r:::::

~~ .-. ~ ~ :;...--

..."..

~ .--

•

C ~

l..",.oI'

.-----

........... ;,.......--

nl'dr SPIi:ld

L.. .....t;:: ~ ~'0

o 20

.40

100

110

1313

'4"0

l~O

1/.00

liO

(~hW~

Figure 4. Xenotest Beta LM - operating areas.

EQUIPMENT OPERATING AREAS (RADIATION - TEMPERATURE) Variable adjustment options and also restrictions required by the linkage of parameters make it absolutely mandatory to provide the user with a clear statement about the possibilities and limits oftoday's equipment systems. For the essential parameters such as radiation and temperature, this can be done in form of equipment operating areas according to the interrelationship characterized in Figure 2. This is based on the evidence gained in numerous measurements in various types of equipment that the temperature difference (BST - CT) as a function of the UV irradiance and the flow rate are independent of the temperature level and the air hunlidity. The corresponding graphs show as a first step general setting options and limits. Figure 4 shows this by the example of the Xenotest Beta LM system and the inserted XENOCHROME filter systems. The requirements specified by DIN 75 202 (Revision: filter system XENQCHROME 320) and SAE 12212 (Filter system XENOCHROME 300) for the state of equilibrium are indicated. Also limits of UV-irradiance can be increased without changing the temperature difference. Figure 5 shows the applicable operating areas for the equipn1ent Xenotest Alpha LM and Xenotest Alpha LM High Energy with filter system XENOCHROME 320 (the corresponding operating area of the Xenotest Beta LM is also shown for comparison). The large adjustment range of the UV irradiance at unchanged temperature difference becomes obvious.

New Equipment Technologies

- - NP'"!OltilX320 60

.. - r> • £'efOX32O - - AJp"tOKE X320

~ 40

X ()

i3 so 20

0 0

t .,

w""

100

L10

l~O

tiO

Figure 5. Xenotest Alpha LM High Energy - Operating areas.

At a low specific temperature level it must be checked in an additional step whether the necessary chamber temperature can be realized - in conjunction with the ambient conditions in the laboratory - or whether additional measures are needed (e.g., utilization of a cooling aggregate).

MEASUREMENTS WITH STANDARD REFERENCE MATERIALS In connection with standards discussion, we already mentioned the application of SRM's for the purpose of detennining whether the xenon-arc apparatus is operating properly. Blue wool light fastness standards as reference fabrics and a clear polystyrene standard as a reference plastic are in use at the present time. Based on the arrangement of the measuring systems for radiation and temperature which are even in a fixed point when positioned on the rotating sample plane, the spatial uniformity can at any time be verified by means of the standards. Figure 6 shows as an example the unifonnity measurement for Xenotest Beta LM with the blue wool reference standard type 6 (~E*, CIELAB D65/100) and polystyrene chips (~b*, CIELAB D65/100) over the length of the specimen holder.

Weathering of Plastics

5.5

Delta b* Polystyrene chips

4 PS-chip no. 125 (ro. 1 top p:GitiorV

Delta b* 7.) Blue wool re erencetype 6

---..

"- -

-1 5

-10

Uf¥ r FOsition

-5

middle FOsltion

Figure 6. Xenotest Beta LM - Unifomlity measurements.

lov~ r FOsition

Apart from the utilization of the SRM's to check whether the equipment is working properly, these "defined materials/material properties" can be used to investigate the effect of selectively changed stress conditions. In a variety of measurements in different systems under variable stress conditions, a basically linear function was found for the SRM's blue wool reference standard type 6 and polystyrene between the property change (8E* or 8b*) and the time (r~0.98). For the quantitative con1parison oftest results the exact definition ofthe kind of exposure of the SRM's and the measuring conditions has to be allowed for. Figure 7 shows an example of how the exposure influences the alteration of the polystyrene standard. Due to the influence of the back reflection of the radiation, the specimen positioned on the stainless steel shows clearly higher ~b* values than the

translucent specimen. Figure 8 shows the effect of an explicitly extreme modification of the spectral energy distribution in the UV range (use ofdifferent XENOCHROME filter systems, Figure 1) at unchanged UV irradiance (300-400 nm) on the slope of the curve. It can be concluded from these results that at least significant spectral differences (especially the presence of radiation in the range below 300 nm) at an identical UV irradiance can be detected by means of SRM measurements. This also includes the risk, which is not unusual in practical application, ofusing the wrong filter systen1. The application of high UV irradiance as a n1easure to accelerate photochemical ageing processes presumes a linear correlation between the observed property change and the applied exposure. A selective evaluation of this measure can only be made when the thennal conditions are unchanged.

New Equipment Technologies

..-

..... ~8£lA·t:no

I--

-.-..

.. • .... ,.

ct'lC"~

"1~Ct'S..11'.f1.. - ., 2QO

-....0

-AlCN.LJd.• '~ "~LM

2.5

.~ ~ V..,... /~~:,....-~ ~- SOO .• ~/ ~ ~ t::z 6D Win' /~ ~ -'" ~ . . / r 65· C ./M ¥~ ~. .. ~ . CH1_ 3&·C .. ~~ ~

- - - 0 - ~ LM.~~a""".", r--

I~ ~

CPS.• 10%

. .. .. ., -~ ~I

~ g ~ A · ~ M•.....",,1Ml _._,~CP'S

4\1,$

..... ~

V ........ ~

0.5

~ ~

r.h.c10'1

(1"/

102~.

dr, '$ mn.

o o

100

'60

.80

'200

'J\~'

Figure 7. Xenotest Beta LM - Polystyrene - ~b* values vs. exposure time.

The corresponding tests were carried out in the Xenotest Beta LM at a limited increase (filter system XENOCHROME 320) and in the Xenotest Alpha LM/Alpha LM High Energy at widely variable UV irradiance (filter system XENOCHROME 300) . The results compiled in Figures 9 to 11 indicate that this precondition is largely fulfilled for the tested materials/properties. However, this cannot automatically be transferred to any other materials/properties. The utilization ofdifferent equipment systems always entails the risk ofdiffering test results due to differences in details, even despite the fOffi1al uniformity of exposure conditions. The proof of identical changes of the SRM reduces this risk. Figure 7 gives a comparison between the linear regression curve calculated from numerous measurements in Xenotest 1200 CPS and Xenotest Alpha LM according to the conditions of ISO 4892-2 (method A) for the polystyrene standard (in the Xenotest Alpha LM apparatus the standard was for certain reasons exposed on stainless steel) and corresponding measurement results obtained in the Xenotest Beta LM. The approximate correlation of the curve slopes (within a tolerance band of ±1 0 %) provides, apat1 from a constancy of the physical measuring and control data, an additional security regarding the largely coincident stress distribution on the exposure plane.

Weathering of Plastics

'--;::=:=:=:::l::======::::::::=========~-I----"""-----I

IXtnotett 8e~1 LM dltfetert fUtMtYt1emr. a,U. wool t.eteMe tyPe t

- +_-----_+_------.. . .-----....__.o1

10 ......- - - -.......--tt--------+---~~ ...........

a ......

........--.....-..-.-~;,------I

-----~I-----........-----.,,.jlIIIi---"----+----------+""'2IIJItfJII:O...o...-

---+_----__4

+----..--st1'..-,i-----~....._.:~-~.....

- I ' 2'0 ~ ~ ~.) ~1C'ttO

~~~.)

2~~~-~~~~~----~------~----~-·~~~~~~~~~~

0..--------.-......-......-..1i-------+-------+-------+------.....-------( o

1:20

100

,rna.. tw:».

~~=:::::===:::::%:==~~==:!::=::::T--1-==:c:==:::c====-.-..-__, X'.:rotl~ BIliLa r..M. dlffttanr FIt.~Mont... P:)~9YMlr.~

- - ~ ~19 ~

~~---+----1

---_+_---.. . _.. .( _.-

o:u.-a

' ' ' 'H .~~ \.flpIoQQ.~

oX ~~ ~

A4NW I.)IC,'I~ J.Not'Clll

.-.0-.. Ie :UIO

....

& -I-----I-----+-------~---_...oi!

,+----+----.....

-------.......--~-..

....

. ....-----t-'........- - - - _ t _ - - " " " - _ t _ - - - - ' " ' t _.......................~--........-__t

6 ......- - -......- - - - - t - - - - 7 " ; ; - - - - - - - - - f - - - - - ; - - - - - t - - - - - t - - - - - - t - - - - - f

of-----+--------.. . ..... .... ........-----.. . ----M.. . . . . . . . .- --.. . .---..-c ----+--~

.. - I - - -........+--.........

~-+--O----~~--+------

~----+-----+---....,~

-----~

2-.1----...,,:..+----........-4----~~ ....."---~.......-.;.....- - -.....- - -......- - - -...---~-----r

.. .. - - - - .--..------:f----

o ......... ...---+--........"'--+-...........u

-!"------!------I

+-------f----+----....--.. . . . . .------+----.. . . .

-'"""'---~

.........

luu

Figure 8. Xenotest Beta LM. ilE* values vs. exposure time. Operating conditions: E(UY) = 60 W RH= 30%.

2 In- ,

CT = 65°C, BST = 90°C,

New Equipment Technologies

Btut w~

,#I.~

tYP4t 6

r--------- -£.~w.

- - 0 - - ~ .. "6r)WIIqm

r - - ----IMII'

,.~_~

-/ /: 3~

~ ~

a • &SJ

66·C

1OO·C

~. 30~

tol InMJJmI

Figure 9. Xenotest Beta LM. Blue wool reference type 6. L\E* values vs. radiant exposure.

--f.

60W~qn

-....0----

E.. 80 Wftqn

_.-

II~

t9;1s'alcn

......

6-+----===:::::;::=======i~---+-- ~~!::.-------.j~---.-..l

- f-----.. . . . . . . . . . . . . . " ""----+------.I

3-+-------+--'-.........'7fIj~-+---- ...........

---+------!

.......

2-+------::iI~.----+-----.-.f---- 4 - -.........

o-t------t------+------+-----.....-----I------~ o

Figure 10. Xenotest Beta LM Polystyrene (on stainless steel) - L\b* values vs. radiant exposure.

Weathering of Plastics

5+----f-----ll-----+----+-----t--.......---+-...........--.-...-+-.,..e;;;;....-~~-......;,.....::;;;;;...-......j

lOOW.&q"rI /Al(;fvJt-l'lt\&"ef1/

:ltOW"'CfI'l J«J'ttJt-i~e-.", --~W/£CI1l AI~lU(rOli1~~

--O-lOS

H .:)00-.4 0C 1m) In MJfn2

Figure 11. Xenotest Alpha LM High Energy - Blue wool ref type 6 - ~E* values vs. radiant exposure.

REFERENCES 1 2 3

Technical report CIE Publ. No. 85; Solar spectral irradiance, table 4; 1st edition 1989. eIE Publ.No. 20 (1972). J. Boxhammer, D. Kockott, and P. Trubiroha, Black Standard Thennometer - Temperature measurement ofpolYlner surfaces during weathering tests, Materialpriifung, 35 (1993) 5, p. 143 to 147. 4 1. Boxhammer, A New Temperature Sensor for Improved Reproducibility in Weathering Tests Handbook, 77th Annual IFAI Convention Orlando, Florida 1989. 5 B. M. Reagan, Accelerated Lightfastness Testing of Disperse Dyes on Polyester Autolnotive Fabrics, Midwest Section's ITPC Committee; 1993 Intersectional Technical Paper Competitions, Dec 1993; Vol. 25, No. 12 (1993), p. 25 to 32. 6 L. A.Bard, USA Automotive Interior Material Testing; Yesterday, Today and Tomorrow, 1993 IPC-7 SAE Conference for Pacific Rim Countries, Phoenix, AZ , Nov 1993. 7 Y. Watanabe and 1. Matsuoka, Study for the Good Correlation and Acceleration on the Artificial Fading Test Methods for Auton10tive Interior Fabrics, International Sylnposium on Automotive Test Procedures for Interior Trin1 Materials; Williamsburg Virginia, August 17-18, 1989. 8 L. A. Bard, Development Program for new SAE Standard J 2019 "Accelerated Exposure ofAutomobile Exterior Materials Using a Controlled Irradiance Air-Cooled Xenon-Arc Apparatus", FSCT - Federation of Societies For Coating Technology 1994. 9 L. Crewdson, A Comparison ofExperimental High Irradiance and Standard SAE Weathering tests for Automotive Exterior Materials, SAE Technical Paper Series 940855; SAE International Congress & Exposition, Detroit, Michigan, March 1994. lOR. Luger, New Possibilities in Filter Technology for Xenon Light and Weather Fastness Testing Equiplnent, Material Life Society; 1st International Symposium on Weatherability, Tokyo, May 12-13,1992.

Weatherability of Vinyl and Other Plastics

James W. Summers and Elvira B. Rabinovitch The Geon C011lpany, One Geon Cenlel; Avon Lake, OH 44012, USA

INTRODUCTION The mechanism of weathering often is caused by damaging radiation from the sun. One way to rate the outdoor light stability of various polymers is to establish the energy to break its bonds. We have done this for a variety of polymers including polyethers, ABS, PCIABS, HIPS, PS, PVC, CPE, CPVC, PP, LLDPE, HDPE, Nylon. Other polymers such as PC, Nylon, and acrylics are sometimes damaged by moisture through hydrolysis. This understanding of degradation can predict differences in color stability and impact retention in dry-sunny climates vs. moist-less sunny climates. Outdoor weathering data is presented for a variety of polymers. PVC performs well in both color retention and impact retention compared to other polymers in these clinlates.

EXPERIMENTAL Bond energies are used fronl the literature for the comparisons of various polynlers. Weathering exposure was carried out according to ASTM D1435 on racks facing 45° south. Materials used are from a variety of sources including Geon rigid PVC, Geon semirigid cap, Korad acrylic cap, Abson ABS (now obsolete), Rovel SAN/EPDM, Geloy ASA, and Plexiglas PMMA. Color change was measured using a Hunter D54 spectrophotometer and is reported in Hunter Lab units. The dropped dart method used was ASTM D4226, H.25 dart, method A.

Weathering of Plastics

400 r-v-ac-u-um----,......--v~i$~ib,...la--,......-ne-a-r...., uilrav,olet

DISCUSSION

mfrared

SUN'S ENERGY SPECTRUM 300

Some of the sun's energy penetrates our atmosphere and has been measured as a function of wavelength,I,2 This is plotted in Figure I, The region of vacuum ultraviolet and ultraviolet wavelengths shorter than 0.28 /lm are absorbed by the oxygen and nitrogen in the atmosphere. If it were not for our atmosphere, the sun would do tenible damage on earth with its ultraviolet energy. Some of the ultraviolet energy, a considerable amount of visible energy, and some infrared energy gets through the atmosphere,

100

O'------"-----o 02

....J

Light wavlll~ngth, micrometers

Figure I. The Sun's spectrum noon, August, Connecticut

010 015 0.20 vacuum uitraviolet

025 030 035 ultraviolet

040

wave lenglh, miCrometers 045 0.50 0,60 0.70 0.80 visible

090 1.0 1.5 near infrared

3q~i

I~ared

r--------- electronic energies - - - - - - - - - . . , ~

f----;.....- bond breaking energies

-(C"OHC"O)-, bela-dikelone eleclronic, 0.43 ·N=N·, ala electronic, 0.35

~- Vibrational energies-

C-H streich, 3451 O-H & N-H streich. 295 t

10205, ester electronic, -(C=O)-O-

10 20, carbon-carbon double bond breaking,

-CH,=CH,-

I 0.145, carbon.;:arbon triple bond breaking, CH "CH

t0.28, carbon-hydrogen bond breaking, CH~-H peroxide bond breaking, HO-OH, 0 53 t

t0.33, carbon-carbon bond breaking, CH -CH 3

Figure 2. The bond breaking energies compared to electonic and vibrational energies.

Figure 2 shows the effect of the spectra of energies on some chemicals. The vacuum ultraviolet causes chemical damage and is of concern for polymeric applications in space,

Weatherability of Vinyl and Other Plastics

where there is little atmosphere for protection. The infrared region has energies that cause molecular vibrations including stretching, bending, or rotating, or common heating ofmaterials without chemical damage. 3 There is a wide region of the spectra from ultraviolet to visible to near infrared that can excite certain chemical structures electronically, from ground state to a higher electronically excited state. While the bonds are weakened in the electronically excited state, normally this is without chemical damage. 4,5 However it is the ultraviolet spectra that has energies capable of breaking chemical bonds. When energies are absorbed, they are capable of breaking chemical bonds. These energies extend into the visible region for son1e weak structures with easily broken bonds such as peroxides.

BOND BREAKING DAMAGE Certain molecules undergo bond-dissociation when an electron is excited, and break into free radicals.

R-X

-+ R' + X'

The energy associated with this event is called the bond-dissociation energy. Also radical attack on a n101ecule ran lead to bond breaking. These bond dissociation energies have been determined for many types ofbonds. 6- lo The relationship between bond dissociation energy and wavelength of light is calculated with the following formula:

A =hC / E where A is the wavelength of light, h is Planck's constant, C is the speed of light, and E is the bond-dissociation energy. And this data can be used to compare the inherent stability ofvarious polymers to damaging energy from the sun. Figure 3 compiles various chemical structures and their bond dissociation energies and lists corresponding polYlners which contain these chen1ical structures. As detem1ined by this table, polymers can be ranked according to the weakness of their bonds. The weakest polymers would be polyethers, ABS, PCI ABS, polymers containing MBS and MABS, HIPS, PS, ASA, and SAN/EPDM. Next, polymers with stronger bonds would be PVC, CPE, CPVC, PP, LLDPE, HDPE and Nylon. Those polymers generally having strong bonds for ultraviolet light resistance, but weak in terms of poor hydrolysis resistance would be PC, PMMA, acrylics, polyesters, and PET. When bonds break in polymers, molecular weight is reduced, leading to brittleness. Loss of itnpact is often a good indication of the degradation that takes place. Loss of impact can also involve physical aging (annealing), so sometimes other measurements are necessary to separate chemical damage from physical aging. 14 Some polymers containing esters, amides,

Weathering of Plastics

Model compound's structure

Bond-dissociation energy, Kcal/mole

reference

Corresponding polymers

very unstable

Polyelhers

61.5

8,9

ASS, PC/ASS, MBS, MASS, HIPS PS, HIPS. ABS, ABA,

CH3-CH2-Q-CH-CH3

~ H

-1-CH3

CH2=CH-CH2

SAN/EPDM, PC/ABS. MBS

CH~CH-CH3

+CI

HOPE, Nylon

PVC. CPE, CPVC

PVC, CPE, CPVC

HOPE

103

114

AN, ABS, ASA, SAN/EPDM, PC/ABS AN, ASS, ASA, SANJEPDM,

PVC, CPE, CPVC

CH3

PP LLDPE

CH3-C.CH3

+H CH3""CH2i-CHz-CH3 CH2CI-f..CH3 CH~CCI-CH3

CH3-CH-CH3 ~

H CH3-CH-CH3

~ H CH3-f-C5N

-Ct

CH3

PC/ABS stable to light, carbonate unstable to hydrolysis PC -(C=O)-NH-, amides stable to light, 12 Nylon unstable to hydrolysis -(C=O)-O-. esters PMMA, Acrylics, stable to light, 13 unstable to hydrolysis Polyesters, PET code: ASS is acrylonitrile/butadiene/styrene, PC/ABS is polycarbonate/ABS alloy, MBS is methyl methacrytatelbutadiene/styrene, MASS is methyl methacrylate/acrylonitrilel butad;ene/~yrene, HIPS is high impact polystyrene,

Figure 3. The inherent stability of polymers based on model compounds.

Weatherability of Vinyl and Other Plastics

II!

or carbonates, loose molecular

5 ...---------.•- - - - - ,

<'!l

--o.-ifl Arizona,

.... 3

,. 0 ~ ,in Arilona,

~ 4 0> C

1l,ay vinyl cap

t weight by hydrolysis due to 1 f

!)i8)' 3CryUC cap'

-+-iIl01110. I gray .my 1(Alp .••. in Ol1io.

~ 1

, __ ._

(.)

~oy acry5cc~Pi

moisture and can become brittle in moist atmospheres. IMPACT AND COLOR RETENTION

It is informative to compare vi-

nyl to acrylic in a sunny, dry climate (Arizona) and in a less sunny, humid climate (Ohio). Vinyl is damaged through 3 =-:------------.., "5 ~---..')---""-{]o.-~ ----o vinyl cap bond-dissociation by ultraviolet i on vinyl i light, while acrylic is damaged !;;'E ;: '· •. gray I t::..E: I , acry1iccapl through hydrolysis by moisture. ~~ 1.5 ~'•...•. ".' ..•.. <."' •• 4. 4,. ! 0f1 vinyl ! Figure 4 shows color shifts in Arl'S 1 ~ gray i acrylic cap! izona and Ohio. Vinyl has its ,---'-lIo-_..............- . ................- . ........._ ................ _ _on.:.c:.:..;ABS : largest change, as expected, in o 12 24 60 Arizona and its major shift is in VVeatherlrrg nmths in orne yellowing then bleaching caused Figure 5. Impact resistance and impact retention of vinyl on vinyl, acrylic on by ultraviolet light, 10 The acrylic, vinyl, and acrylic on ASS (ASTM D4226, h.25 dart). as expected, has its largest change in Ohio and its major color shift is in lightening caused by moisture and hydrolysis of the ester. Impact resistance is another important characteristic in weathering. It varies considerably from material to material. In the Izod test, crack propagation is the main energy measured using pre-notched specimens. In dropped dart impact, initiation of the crack is usually the main energy measured. Vinyl is normally considered ductile; for example, added impact modifier will jump Izod fractures from brittle to ductile and its tensile elongation prior to rupture is usually over 100%. ABS and acrylic (impact modified PMMA) would be considered brittle; for example, added impact modifier will normally not jump Izod fractures to ductile and its tensile elongation is usually less than 50% to rupture. As an example of these ductile and brittle characteristics on weathered samples, Figure 5 shows impact resistance of ductile on ductile materials (vinyl on vinyl), brittle on ductile materials (acrylic on vinyl), and brittle on brittle materials (acrylic on ABS). Another example, for uncapped PVC and ABS, both highly protected from ultraviolet light with 15 parts titanium dioxide per hundred resin, is shown in Figure 6. Both the initial lack ofductility and the rapid degradation ofthe ABS are evident in the plots. Here the ABS is We3'heri'lg monthS

Figure 4. Color shift of vinyl and acrylic in sunny-dry (Arizona) and less sunnyhumid (Ohio) climates. •

i~·gray

l o~

_ , . , . , ....

. . ~ l

Weathering of Plastics

3r----------------.

f ........... pvc

Arizona

.. o·ABS

MZ0i18

1.5

-6-PVC Ohio

·· .. ··ASS Ohio

o <--------~.!t_'"~.'-'-.'-'...;;.;''c:.'""-".:.;''.;;.;":..:.'" ,--_--, o 6 12 16 24 Outdoor weathering. morths

Figure 6. Impact resistance and retention ofPVS and ABS, 15 phr Ti0 2 (ASTM 04226, method A, dart h.25).

~ --"9io

.~,~~~ i-~~ ~

....., . , . , .... ..• _

• . • . /r' ... - ' - - " ' ' ' ' - ......

! - ..... ·SAAlIEPDMI

l' •.. PM".;" ...• -

gerrlM~c

File

- r.. - ecr,lle

pi=ln~

24 Weat~e,

agirg in Anzona, mont.,!,

Figure 7. Color weatherability of white materials in Arizona.

.---------------.-

nO'd

p'J(,

..... ,JO . . . . .

.,'

....

--o--A.i,jA

,If

....

-4I-Mrniroi:.f P";}

. ° .~~~;~':~!l::".~~-;~~_\ -~ _ _-'--_ _..... Oc:..:._ _...Jx-:..:..:.:..~_ _'___ _ 0

c~tII~ro'"

... ·

-i}. •..•

u~lit:

P'l~~

I J

less ductile than the PVC, showing lower impact initially, and the ABS structure, particularly the polybutadiene, is weak and susceptible to the damage and oxidation caused by the sun, thus resulting in poor impact retention. In Figures 7 and 8 various white materials are examined in both the sunny-dry climate (Arizona) and the less sunny-moist climate (Ohio). In both Arizona and Ohio, over a five year period, the white rigid PVC changes little in color, similar to PMMA, semirigid PVC cap, and acrylic paint, and performs better than SAN/EPDM and ASA polymers. Brown materials are compared for color change over a five year period, also in Arizona and Ohio, and the data are presented in Figures 9 and 10. This data shows the ASA has more color change than rigid PVC, particularly during the first three years. At five years of weathering, and especially in Ohio, PMMA changes in color more than rigid PVc. MICROBIAL ATTACK

Figure 8. Color weatherability of white materials in Ohio.

Another comparison of plastic materials concerns resistance to microbial attack. PVC has high resistance to microbial attack, a reason for its success in windows and siding. The ASA plastic shows poor resistance to microbial attack and several cleaners were unsuccessful in eliminating the microbial stains. 15

Weatherability of Vinyl and Other Plastics

Table 1. Buckling resistance of plastics in a window profile 16 Coefficient of linear thermal expansion

Material

em/em °C

in/in of

Rigid PVC

2.1 x IO- s

3.7xlO- s

High heat PVC

1.9x IO- s

Safety margin (additional stress to cause buckling)

Heat deflection temperature, °C

white

brown

MPa

psi

MPa

psi

6.0

870

3.5

510

3.5xlO-s

6.9

1000

5.0

730

ASA

2.7xlO- s

4.9xlO-s

4.0

580

2.0

290

Glass fiber reinforced PVC

0.9xlO-s

1.7xlO-s

2600

2180

3,5

" :t.

,-tlgid

PIle '-<:r--/,SA

•/

~r', 'N"1;.~.:"..,,_:~~

;.;,r

,'" ~'

1/("

,.,. III -

'''-...

$emirig'id

PVC

"..-

. <)0.

acrylic pain1

Figure 9. Color weatherability of brown materials in Arizona.

-ti(Jkl

I'VC , -e-ASA •

-" .. () ..

/,,-w.

serrririQiO

pvc

~Cty'l¢

pain1

o ---------.------o

PMMA

Outdoor agi1

Figure 10. Color weatherability of brown materials in Ohio.

DIMENSIONAL STABILITY AND WARPAGE

Wrinkling, "oil-canning", and warpage are fonns of buckling. In the sun, plastic parts are heated directly in the sun more than the areas not exposed to the sun. This causes differences in thennal expansion and puts the shaded areas in tension and the areas exposed to direct sunlight in compression. Buckling can be calculated using mechanical engineering equations and depends on the compressive forces on the material and the design of the structure. For example, Table 1 shows the safety margin resisting buckling in a typical brown plastic window in full sun. 16 This data shows why the cooler white window resists buckling better than a brown window. It also

Weathering of Plastics

shows the main factor for buckling is coefficient of linear thermal expansion and not heat deflection temperature. Also, the ASA plastic window would buckle with the addition of 290 psi more stress while it would take 507 psi to buckle the PVC plastic window, thus the PVC is more resistant to buckling than the ASA.

CONCLUSIONS To summarize, much of various polymers' weatherability is predicted by the strength of its chemical bonds, either broken by ultraviolet light or hydrolyzed by moisture. Good compounding technology certainly is also required. PVC usually has superior microbial resistance. Buckling resistance is better with a low coefficient of linear thermal expansion, where PVC often has the advantage. When rigid PVC and semirigid PVC capping materials are compared to other polyn1ers, the PVC performs well as predicted and compares favorably to other polymers.

ACKNOWLEDGMENT We wish to thank the Geon Company for pennission to publish this paper. Thanks to Bill Northcott for maintaining our weathering infoffi1ation.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

R.C. Hirt, N.Z. Searle, Applied Polymer Symposia, No.4, 61 (1967). V. Schafer, Applied Polymer Symposia, No.4, 111 (1967). R.M. Silverstein, G.C. Bassler, Spectroscopic Identification of Organic Compounds, p. 56, Johll JViley & SOilS, New York (1963). J.W. Summers, Electrically Semiconducting Polymers, Ph.D. thesis, p. 2, Case Western Reserve University (1971). R.M. Silverstein, G.C. Bassler, Spectroscopic Identification of Organic Compounds, p. 91, John JViley & Sons, New York (1963). J. Hine, Physical Organic Chemistry, 2 nd Edition, p. 422, McGralv-Hill Book Company, Inc. (1962). H. H.G. Jellinek, Applied Polymer Symposia, No.4, 54 (1967). C. Walling, Free Radicals in Solution, p. 48-50, John Wiley & Sons, Inc. New York (1957). G.M. Badger, "Pyrolysis of Hydrocarbons", Progress in Physical Organic Chemistry, Vol. 3, p. 4, Interscience Publishers, New York (1965). James W. Summers, Elvira B. Rabinovitch, J Vinyl Techn., 5 No.3, 91 (1983). 1. Hine, Physical Organic Chenlistry, 2 nd Edition, p. 32, McGrawHill Book Company, Inc. (1962). Anon, Plast. Mod. Elastomers, 21, No.II0, 81 (1969). C. Walling, Free Radicals in Solution, p. 412, John Wiley & Sons, Inc, New York (1957). Elvira B. Rabinovitch, James W. Summers, William E. Northcott, J Vinyl Technol., 15,4,214 (1993). A study done and reported by the Vinyl Window and Door Institute of the Society of Plastics Industry. Elvira B. Rabinovitch, J Vinyl Technol., 10, No.1, 14 (1988).

Aging Conditions' Effect on UV Durability

Robert L. Gray, Robert E. Lee, and Brent M. Sanders Great Lakes Che111ical Corporation, P.O. Box 2200, West LafaJ)ette, Indiana 47906 USA

INTRODUCTION Efforts to carry out intermaterial substitution in thermoplastic resin industries continue to be driven by economic advantages and perfonnance enhancements.! This has forced producers to document performance attributes such as UV durability in an effol1 to make overall part-life predictions. Participating aggressively in the areas of UV stabilization and accelerated aging has provided us key insights to how factors of aging impact performance. While it is vital to recognize that specific testing or aging conditions have the potential to greatly alter performance measurements, the understanding of mechanistic causes for these phenomena can be a tool to breakthrough performance technology.

BACKGROUND Plastics have replaced wood, metal and other materials of construction in many applications. Because of the combination of strong economics and perfonllance enhancements, some applications appeared uniquely suited to a particular type of plastic resin. For example, "Vinyl Siding" and "Polyester or Nylon Carpet" have captured large market segments by displacing natural materials like wood and wool, respectively. Although these types of substitutions still take place, a tremendous amount of work is now being done to carry out intermaterial substitutions. This is where polyolefins replace higher cost resins and resins of different manufacturing methods (metallocene vs. conventional catalysis) are competing within resin types. Previously, part performance was so drastically different that direct comparisons where difficult. However, with intennaterial substitutions, con1parison is both possible and required. Time-to-market and product approval constraints have thus made accelerated aging a common practice. 2 However, there has been a near constant debate over the correlation between test methods. For example, cOlnparisons of oxidative induction time (OIT) experiments to long term

Weathering of Plastics

heat aging (LTHA) illustrate the large differences accelerated testing methods make on predicting thermal oxidative durability.3 This debate is fueled by the large number of factors affecting the service life as noted in Table 1. 4 Table 1 showing the degradation factors affecting the service life of building components and Inaterials.

Table 1. Degradation factors that affect building components and materials Radiation

Biological Factors

Solar

Microorganisms

Nuclear

Fungi

Thermal Temperature

Bacteria Stress Factors

Elevated

Stress sustained

Depressed

Stress periodic

Cyclic Water Solid (snow, ice) Liquid (rain, condensation, standing water) Vapor (high relative humidity) Nonnal Air Constituents Oxygen Ozone Carbon dioxide Air Contaminants Gases (oxides of nitrogen and sulfur) Mists (aerosols, salt, acids and alkalies dissolved)

Physical action rain, hail, sleet and snow Physical action of wind Movement due to other factors Incompatibility Factors Chemical Physical Use Factors Design of system Installation and maintenance procedures Nonnal \vear and tear Abuse by the user Particulates (sand, dust, dirt)

MECHANISMS OF STABILIZATION

Because of the large number of factors listed, users are often confined to one particular test method to serve as a bench Inark or an approval criteria. We have compared a few common aging tests for UV durability in our work and focused on factors related most directly to additives used for improving UV durability. Extensive reviews are available on light stabilization and light stabilizers. 5 They indicate that degradation generally occurs by free radical mechanisms that are inhibited by radical traps, excited state quenchers or UV absorbers. Degradation can be observed as changes in color, appearance, strength or elongation. However, these are the results of chemical bond cleavage and the formation of other chemical bonds.

Aging Conditions' Effect on UV Durability

Radical traps of the hindered amine light stabilizer (HALS) type have been extensively 6 used to combat UV degradation. Bis (2,2,6,6-teramethyl-4-piperidyl) sebacate (HALS 1) is a 7 well studied ,8,9 example of this type. The active portion is the 2,2,6,6-teramethyl-4-piperidyl moiety which has been incorporated into a wide variety of molecular architectures for better compatibility and performance. After pre-oxidation of the nitrogen to the corresponding nitroxyl radical, the active moiety undergoes a cyclic mechanism oftransfonning reactive peroxide radicals into alcohol and carbonyl species as shown below. 10 However, this is just stopping the process after damage has OCCUlTed. The use ofUV absorbers (UVA) can prevent damage by dissipation ofUV energy by mechanisms shown below. II

A Cyclic mechanism of radical trapping for HALS.

Mechanism of energy dissipation for UVA's.

MIGRATION AND LOSS OF STABILIZERS

Light stabilizers used in any polymer system must be present and available at the sight ofdegradation. It has been shown that stabilizers can be absorbed into pigments such as carbon black and effectively removed from the system. Another loss mechanism can be by chemical interactions. For example, acidic byproducts of halogenated flame retardants deactivate hindered amine stabilizers l2 with high pKa's. Recently, we have shown that in flame retarded polypropylene fiber loss of stabilizers can be reduced by enhancing HALS' compatibility with polypropylene. This reduces their tendencies to migrate into the polar flame retardant regions of the system. 13 The concept of molecular mobility of stabilizers within the polymer system matrix is key to understanding actual perfOimance. The automotive industry has been aware that rain and especially acid rain reduces UV performance as seen by the responses of the same system in

Weathering of Plastics

different exposure climates. In addition to this, migration of the stabilizers into plastic substrates during baking has been isolated as an additional loss mechanism. 14 However, not all migration ofstabilizers is bad. For example, in a thick part, the light stabilizer is not beneficial below the area of UV degradation. Therefore, its migration renews the surface area where loss and consumption may be occurring. ACCELERATED AGING Automotive engineers have often relied on "outdoor" UV aging at test sites in South Florida, Arizona (desert) and other areas of the world having extremes in climate. However, they also have accelerated methods available which use fluorescent lamps (QUV) or xenon arc lamps (Weather-Ometer). In our labs, SAE J1885 is a standard xenon test method used for automotive customers. In many respects, it represents the peak exposure received at 45° South in "outdoor" testing but with longer light cycle duration. Table 2 shows the key variables for SAE J 1885 and other Xenon test methods.

Table 2. Xenon test conditions Irradiance @340 nm

Black panel,oC 63

63 63 63 55 70

35 35 35 55 50

yes

SAE J1960

0.35 0.35 0.35 0.35 0.3 0.55

SAE J1885

0.55

Test type ASTM G26, Method 1 ASTM G26, Method 2 ASTM G26, Method 3 ASTM G26, Method 4 ASTM D4459

Relative humidity

Dark cycle

no yes

no yes yes

Water spray yes yes

Comments ASTM D2565 AATCC 16E

no no no

Var. of ASTM G26

yes

Spray in light and dark

Unlike automotive users, "Office-Business Machine Cabinets" see much different UV exposures during their service life. Therefore, ASTM D4459 is a xenon accelerated UV aging method often used to simulate peak exposure behind window glass. Note that the irradiance level (Watts per square meter) is substantially higher for the automotive test. This means a higher energy flux. Also, the 34°C increase in the black panel temperature (relative to part surface temperature) is dramatic. In nlany solution phase reactions, a 10°C increase could account for a doubling of a reaction rate. Therefore, degradation reactions should also occur at an accelerated rate on the polYlner surface during UV testing.

Aging Conditions' Effect on UV Durability

EXPERIMENTAL Compounding polypropylene systems was accomplished using either a Berstorff ZE-25 mm twin screw extruder or a Killion single screw extruder. Temperature profiles nominally were 200-230°C except as noted. Fiber extrusion was performed with a fiber line (Hills Research & Development - West Melbourne, FL) with temperature profiles matched to those used in compounding. A 72 round filament spinneret was used to achieve 18dpfPOY fiber with a 3 to I draw ratio fiber. Accelerated UV exposure testing was done with xenon arc under method SAE J1885 (interior automotive) ASTM G26 (wet and dry) and ASTM 0-4459 (interior dry xenon). Sample fibers were wrapped around 6 cm x 15 cm cards and clamped into standard specimen holders. Tensile strengths were measured using a Instron model 1123 with a 200 pound (90.7kg) load cell, 8 cm sample length, 12.5 cm/min pull rate and a 5 pound (2.268kg) calibration weight.

RESULTS AND DISCUSSION Work comparing SAE J 1885 and ASTM 04459 on the same flame so retarded polypropylene fiber 15 ., -.... ... 60 lit .. indicated that the degradation .....y._40 ... ,,, did not correlate to total expo-SAEJHl85 20 sure as measured in kJ/m 2• Spe• ASThf 04459 o L..----------rTT"--'t""~---......J cifically, the loss of tensile o 150 300 k. n 450 600 strength was 0.3% per kJ/m2 for Figure I. Comparisons of xenon test methods for unstabilized flame retarded PP SAE J 1885 and 0.075% per fiber with 6% bromine. kJ/m 2 exposure for the unstabilized system. Data similar to that above has driven some to advocate accelerated analysis of samples aged by non-accelerated methods. 16 A balance should be in place to allow for the use ofaccelerated tests as screening tools. This is especially true when one considers that a year of natural exposure can valY 20% in total radiation from year to year. 17 For any test to be of value it must produce repeatable results. It must also correlate to actual service life or natural weathering in the case of UV testing. The xenon tests in general have very good repeatability. Therefore, one could make an attempt to "scale" results to natural weathering. However, if one would say that 600 kJ/m 2 of ASTM 04459 exposure is equal to 150 kJ/m 2 of SAE J 1885 exposure based on Figure 1, it could be true for the system shown and not others. The way this becomes evident is when one evaluates different systems with both methods and the ratios of degradation per unit of exposure changes. This is particularly 100

.....: -

lit

Weathering of Plastics

harmful when an accelerated test favors a system which is poor in natural aging conditions. Figure 60 -- - -1- - 2 shows the correlation between :+ 40 + a specific Florida exposure se--l.-ries and and the duplicate series 20 +1 I exposed with a xenon technique o ' (ASTM G26). Since the points o 100 200 300 400 Hours of Xenon to reach 70% Tensile do not fall on a straight line there is not perfect correlation for these systems. Therefore, the Figure 2. Correlation between xenon and Florida exposure user must be sensitive to aging factors which could contribute to any lack of correlation. According to Carlo Neri,18 accelerated UV test like SAE J1885 could indicate that HALS 1 outperfonns a polymeric stabilizer like HALS2 because HALS I has a higher rate of migration to the surface of the polymer. However, the reverse perfOlmance is observed for natural weathering because the same high migration rate for HALS I leads to depletion of the stabilizer (removal by rain, etc.) in an untimely manner. The presence of water spray during the xenon testing can highlight differences as shown in the polypropylene systems containing HALS3 as indicated in Table 3. Table 3 compares the irradiation required to reach 50% of the initial tensile strength for fibers containing different HALS. Column 2 contains the results from ASTM G26 exposure with includes water spray. Column 3 shows the results for SAE 11885 without water spray. Column 4 shows the scaling factor or ratio between the methods. The systems shown in Table 3 and Figure 3 represent stabilizers often incorporated into agricultural applications. Therefore, evaluations made using a dly xenon method such as SAE J 1885 could have overestimated the utility ofHALS3 or blends ofHALS2 with HALS3 in the

,...---...,..-----:----r---. ! ~ so .----.1--

100

t--

Table 3. Comparisons of UV exposures with and without water spray [emailprotected]% None HALS2 HALS3 HALS2 & HALS3 HALS4 HALS4-Me

T so % ASTM G26 kJ/m 2

Tso% SAE J1885 kJ/m 2

Ratio G26/J1885

127 1049 318 1Ol3 1318 1196

53 463 431 539 581 555

2.4 2.3 0.7 1.9 2.3 2.2

Aging Conditions' Effect on UV Durability ---44 a--------r:II:::N-:-on:-:-o~

1400 ,.-

1200 - i - - - - - -

IIHAL$2 ,.,.....,.--------IOHAlSJ tlHALS2t~

1000

t--------'lIt1AlS4 DHALS4·Mo

BOO kJ!m2 ~oo

400 200

G26

Figure 3. Xenon test results ASTM G26 and SAE J1885.

J16SS

presence of water spray. The use of HALS3 is promoted because of its low pKa and tertiary amine structure. A methylated version of HALS4 may accomplish this without the sensitivity to water spray. Even the good performance of this HALS compared to the industry "Work Horse" HALS2 is very encouraging for long tenn stabilization of agricultural products.

CONCLUSIONS The factors of irradiation level, part temperature (black panel temperature) and the presence of water spray have been shown to be key aging conditions which affect UV durability. Higher irradiation levels and part temperatures often lead to failure at a lower total light exposure. Likewise, the presence of water spray has been shown to be antagonistic to the performance of certain stabilizers. Overall, the use of accelerated UV aging methods continues to be required. However, care must be taken to correlate all data to real part perfonnance and service life.

ACKNOWLEDGMENTS The contributions and research effort of Olga Kuvshinnikova and Bill Fielding are gratefully acknowledged. Special appreciation is also extended to Great Lakes Chemical Corporation for pennission to use the data presented.

REFERENCES 2 3

4 5 6 7

Singh, B.B.; TPO's - An Industry in Transition, conference proceedings insert FLEXP096, Houston, TX.; June 26-28,1996. Scott, J. L.; Does Correlation Exist Between Accelerated and Conventional Outdoor Exposures?; Atlas Sun Spots: Vol 9; Issue 21,1979. Gray, R. L.; Accelerated Testing Methods For Evaluating Polyolefin Stability, Geosynthetic Testing for Waste Containment Applications, ASTM STP 1081, Robert M. Koerner, editor, American Society for Testing and Materials, Philadelphia, 1990. Scott, J. L.; Does Correlation Exist Between Accelerated and Conventional Outdoor Exposures? Part A Atlas Sun Spots: Vol 10; Issue 23, 1980. Gugumus, F.; Plastic Additives; 4th Edition, editors GachterlMuller; Hanser; I993;Chapter 3. Light Stabilizers. Gugumus, F.; Plastic Additives; 4th Edition, ed. GachterlMuller; Hanser; 1993; pp 190-191. Allen, N.S.; Homer, 1.; Mckellar, J.F: Makromol. Chem., 179 (1978) p. 1575.

8 9 10 II 12 13 14 15 16 17 18

Weathering of Plastics

Ivanov, V.B.; Shlyapintokhl, v.Ya.; Shapiro, A.B.; Khvostach, a.M.; Rozantsev, E.G.; IZl( Akad. Nauk SSSR, Sel: Khim., 8 (1974) p. 1916. Carlsson, OJ.; Grattan, OW.; Supmnchuk, T.; Wiles, O.M.: 1. Appl. Pol)'m. Sci., 22 (1978) p. 2217. Klemchuk et al.; Polymer Degradation and Stabilization; (1988) 22, p. 241; (1990) 27, p. 65. Calvert, J.G.; Pitts, J.NJr.: Photochemistry. Wiley, New York, (1967) p. 535. Klopffer, W: 1. Pozrmer Sciellce, Polymer Symposium, 57 (1976) p. 205. Chadet, J.H.; Newland, G.c.; Patton, H.W; Tamblyn, 1. W; SPE TrailS., 1 (196\) p. 26. Gray, R. L.; Lee, R. E.; "The Influence Of Co-Additive Interactions On Stabilizer Perfonnance" ANTEC '96, May 5-9, (1996), Indianapolis, IN. Gray, R.L.; Lee, R.E.; "UV Stabilization Of Polypropylene FiberContaining Organic Bromine Flame Retardants" ANTEC '96, May 5-9, (1996), Indianapolis, IN. Haacke, G.; Andrawes, F.F.; Campbell, B.H.; Joul'llal ofCoatings Technology, 68, No. 855, (\996) p. 57. Fielding, WR.; Lee, R.E.; "Advances in Flame Retardant UV Stable Polypropylene Fibers", Additives '96, Febmary 18-21, (1996), Houston, TX. Gerlock, J.L.; Atlas School for Natural and Accelerated Weathering, May 11-13, 1994, Miami, FL. Scott, J.L.; Atlas School for Natural and Accelerated Weathering, May 11-13, 1994, Miami, FL. p. 20. Neri, c.; Great Lakes Chemical Italy, private conununications.

t1ALSI

HALS2

H N··-)~) ~--./

HALS)

HO I

··O-(CH2kSi

'---+---11

H HALS"

HALS4·Mc

Molecular Weight Loss and Chemical Changes in Copolyester Sheeting with Outdoor Exposure

D. R. fa*gerburg and M. E. Donelson Eastnlan Chenlical ConlpanJ J, Kingsport, TN, USA

INTRODUCTION We have recently reported on the development of a weatherable copolyester system that employs a coextruded construction ofa layer highly loaded with an effective UV absorber on top of an unmodified copolyester layer that is the bulk of the sheeting. 1 This coex sheet has been found to be very resistant to weathering-induced changes such as color and impact strength. Our latest data indicates good results for 2 yr. exposure in either Miami, Florida or the Arizona desert. In order to gain a more fundamental understanding of the effect of weathering-induced changes, we have done experiments with unstabilized copolyester itself. The literature for polyethyleneterephthalate (PET) film and fiber would indicate that a number of changes are to be expected. 2-6 These involve chain cleavage at numerous points to give a wide range ofproducts such as acid groups, vinyl end-groups, and others and also oxidation ofvarious species, mainly the aromatic rings in the chain to form hydroxylated terephthalate species. Not mentioned in these reports, however, is the possibility of chain branching via reaction with aromatic rings of aryl radicals produced in the photo-oxidation process. In addition, this previous work deals with oriented PET and not an amorphous copolymer.

EXPERIMENTAL Three millimeter thick sheeting of polyethyleneterephthalate modified with 1,4-cyclohexanedimethanol was extruded by the Technical Services Laboratory at Eastman Chemical Company. The sheeting was cut into specimens 30.5 by 30.5 cm (12 by 12 in.) and placed on exposure at the DSET test facility in New River, Arizona. Samples were mounted facing due south with the rack at 45 degrees to the horizontal. A backing sheet, positioned ca. 2 cm behind the exposed plaques, was coex sheeting of the same copolymer with a 5 mil layer

Weathering of Plastics

of the copolymer heavily loaded with an effective UV absorber. Sanlples were exposed for 19, 28, 42, 56, 70 and 83 days. Plaques were cut from samples after exposure were milled with a vertical end mill to depths of 0.25, 0.51, 0.76, 1.52, 2.16 and 2.41 mm (10, 20, 30, 60, 85 and 95 mils) from the exposed surface. An additional set of plaques were cut and were milled to 50, 125 and 250 microns (2, 5 and 10 mil) steps. Micro IR was used to obtain spectra of the exposed and back surfaces and selected step surfaces using attenuated total reflectance (ATR) mode. The spectra obtained were normalized to each other. Areas of peaks were obtained by drawing baselines usually on either side ofa desired peak and not over large areas of the spectrum. Gel permeation chromatography (GPC) was performed on samples by scraping the exposed surface and milled steps along with the back surface of the samples with a sharp knife to a depth of approximately 1 to 2 mils. About 4 mg of each sample was dissolved in methylene chloride-hexafluoroisopropanol azeotrope (ca 70:30 parts by volume) and run on a Perkin Elmer instrument at room temperature through a single PL Gel mixed column. Detection was by UV at 255 nm. Instrument calibration was established with PET standards of knownM w •

RESULTS AND DISCUSSION Space will only permit discussion of a few of the most important IR peaks observed. Examination of the front surface of exposed plaques with exposure time (Figure 1) showed significant changes in several functional groups consistent with oxidation and also chain cleavage. The carbonyl group showed a large increase that was essentially linear with time ofexposure. The primary reason for this increase was that the peak was broadened significantly (Figure 2) which argues for the production of multiple, different species than were originally present in significant concentration in the unexposed polymer. These are the result of oxidation and chain cleavage. Model compounds verify that oxidation of terephthalate units to mono- and dihydroxyterephthalates would produce carbonyl frequencies lower than in terephthalate esters themselves. Additionally, chain cleavage would produce acidic end groups which would absorb at a lower wavenumber than the ester. The OH and CH region rose over a factor of 5 in area which rise was really solely due to broad signals from OH species. Figure 2 shows a comparison for the time zero specimen compared to the 70 days exposed sample to dramatically illustrate this point. The Sp3 CH peak showed a trend downward in area consistent with oxidation of these moieties and hence their diminution. This is, of course, oxidation of the glycol moieties. From model compounds we have been able to determine that the OH groups appearing in the region above the CH peaks can be classed as arising from 2-hydroxyterephthalate and 2,5-dihydroxyterephthalate species and those appearing below the CH peaks as coming from acid end groups. The oxidative processes would seem to show an induction period to 19 days exposure followed by a rapid

Molecular Weight Loss

SO 40

":ij

-e

-e-OH and CH ..... C=O -l>- CH (sp3) - A r ring (730 em·1)

I ,.:~

'"..'"

, 10

,J!!

0 10

Figure 1. IR absorbances of selected bands vs. exposure.

"-OH and CH ..... CH(sp3) -l>-C=O - A r ring (730 em'l)

abs.

20 soecimen deoth. mils

Figure 3. IR peaks vs. depth at 70 days exposure.

'f 'v ~ly~

Figure 2. IR spectra of copolyester at test start (bottom curve) and 70 days exposure (top curve).

Weathering of Plastics

1',210

l1.1 10~

6000

1 10

4000

9000

2000

:;" ~

-e-70 da. e:.<posure --- 83 da exposure

8000

8000-i--......--..,...-.-.,.--~~~:r:----:~----::l~--±

O+------:T~--r-----::T':"O---:~-~~-1:-r40 depth, mils

exposure. days

Figure 5. Number average molecular weight vs. depth.

Figure 4. Number average Inolecular weight vs. exposure.

25 --- front face '-0- 10 mils depth

-&-70 da exposure --- 83 da exposure

MzlMn

~ 15

lJ.

;a 10

exposure, days Figure 6. Mz/Mn vs. exposure time.

o.. . . . . -_--.......,....--~-_-- __- -.. . 20

120

depth, milS

Figure 7. Mz/Mn vs. specimen depth.

the surface. No real evidence for chain scission as measured by M n was observed 250 nlicrons (10 mils) into the plaque. This steep fall-off of damage also means that GPC curves obtained are subject to large errors in sampling as small differences in the depth of scraping for the sample could easily skew results significantly. This can be seen in comparing M n for the 70 day exposure with the first set ofmilled specimens (Figure 4) and the second set (Figure 5).

Molecular Weight Loss

Equally dramatic was the rise in Mz/Mn (indicating branching) (Figure 6) even at short exposure tin1es. This branching reaction is n10st easily understood by reaction ofphenyl radicals produced in the chain cleavage process reacting with terephthalate rings of neighboring copolyester chains producing biphenyltricarboxylic acid moieties. There is a potential slight indication ofbranching at 10 mils into the plaque (Figure 7) at long exposure time. It is not entirely clear, however, whether in fact this might just be randOlTI variation of the test results. One fact is certain, however, that any damage seen at 10 mils depth even at the longest exposure was very slTIall indeed.

CONCLUSIONS The IR and GPe data taken together showed that weathering dan1age to the our amorphous copolyester system was confined to no more than 10 mils of material and, therefore, that the bulk of the plaque was completely unaffected. The changes in the surface showed a great deal of oxidation and also chain cleavage. Oxidation to produce hydroxyterephthalates appeared to have an induction period. Branching of the san1ple rose dramatically with exposure as well. This is consistent with some of the chain cleavage products being phenyl radicals capable of attack on aromatic rings to form branching species.

ACKNOWLEDGMENTS The authors would like to acknowledge the able assistance of many coworkers, among whom are Helen Richards, Brenda Forbes, Edgar Gamble, Joda Wood, Stewart Millen and George Caflisch.

REFERENCES 1. 2. 3. 4. 5. 6.

D. R. fa*gerburg and M. E. Donelson, J. Viny! and Additive Techno!., 3, 179 (1997). J. G. Pacifici and 1. M. Straley, PO~l'm. Lett., 7,7 (1969). G. Valk, M.-L. Kehren, 1. Daamen, Angelt: Malo'om. Chemie, 13,97 (1970). M. Day and. D. M. Wiles, J. Appl. Polym. Sci., 16,203 (1972). P. Blais, M. Day and D. M. Wiles, J. Appl. Po~rm. Sci., 17,1895 (1973). For a synopsis see for example "Polymer Photodegradation, Mechanisms and Experimental Methods", Jan F. Rabek, Chapman and Hall, London, p. 288-94 and references cited therein.

Fourier Transform Infrared Micro Spectroscopy.

Mapping Studies of Weathered PVC Capstock Type Formulations. II: Outdoor Weathering in Pennsylvania

Dana Garcia and Janine Black ElfAtoche111 NA, 900 First Avenue, King ofPrussia, PA 19406, USA

INTRODUCTION Siding and window profile are two major applications of polyvinylchloride (PVC) requiring good long term weatherability. PVC, like other industrially important polymers, is susceptible to degradation under a variety of external conditions such as heat, light, oxygen and mechanical stress. All these factors are present in outdoor applications. It is then not surprising that significant effort 1-4 has been directed towards understanding the mechanistic pathways of degradation, the effects ofexternal factors and the influence ofpolymer structural defects and formulation additives. Fourier Transform Infrared Spectroscopy (FTIR) coupled with microscopy provides unique opportunities to study change, based on the complexity and specificity of the infrared spectrum and the dimensional resolution of the microscope. Jean-Luc Gardette and co-workerS- Il have published a number of papers related to FTIR studies of PVC photodegradation. Photodegradation was investigated by monitoring the thickness direction profile of the absorbance centered at 1730 cm- I. The absorbance is characteristic ofC==O containing structures. Among the species identified were the beta chlorocarboxylic acid (1718 em-I), the acid chloride (1785 cm- I), alpha dichloroketone (1745 cm- I), isolated and conjugated polyenes (1650 cln- 1 and 1580 cm- I). The resulting profiles exhibited the expected decrease in c==o species concentration from the irradiated surfaces into the sample center. The minimum was observed at approximately 200 microns. The production of these species requires oxygen, thus oxygen diffusion was postulated as the controlling kinetic step. Addition of TiO2 (2 to 16%) does not alter the species characteristic pattern of the IR spectrum as

Weathering of Plastics

compared to the unpigmented pvc. The presence ofTi0 2 provides a photo protection effect. Monitoring the evolution of the 1718 cm- 1 band as a function of irradiation time, the authors found a decrease of oxidation with increasing pigment content. The largest difference is observed between unpigmented and the lowest pigmented sample. Similar studies were reported by C. Anton-Prinet. 12 The accelerated photo degradation, of formulated Ca/Zn stabilized PVC, was monitored in the IR via changes in the absorbance ofthe c==o region, 1720 to 1740 cm- 1• The depth of oxidation products was determined at 100 to 200 n1icrons from the light exposed surface. The carboxylate absorbances characteristic of the calcium stearate and zinc stearate stabilizers were observed to decrease in the layers close to the exposed surface as stabilizer was consumed. No attempt was made to separate the various species responsible for the absorbances in the c==o region and determine individual profiles. Previous results from this laboratolyl3 examined the QUV accelerated photo degradation of PVC capstock formulations with varying levels ofTi0 2 • The most pronounced effect observed via FTIR micro spectroscopy mapping techniques is the loss of the calciun1 stearate carboxylate absorbances. The effect is attributed to the reaction of calcium stearate with HCl to form CaCl 2 and stearic acid. 14 The presence ofCa in the layers close to the exposed surface was continned by ICP (inductive coupled plasma) measurements, indicating that there is no physical loss of the material. The reaction of the calcium stearate proceeds the furthest, from the exposed surface, into the PVC sample and is dependent on Ti0 2 1evel. The behavior ofthe 1780 cm- l and 1718 cm- l absorbances lag behind the reaction of calcium stearate. The 1718 cln- l absorbance appears earlier in the process and proceeds to a greater depth than the acid chloride. The behavior was explained by the fast reaction, as it is being formed, of the acid chloride to the chlorocarboxylic acid. The presence ofTi0 2 significantly reduces the level of photodegradation. Both the concentration of degradation species and the depth of degradation are significantly decreased. This paper continues in the same direction by providing a systematic FTIR study of changes observed during outdoor photooxidation. The objective is not only to simply obtain the depth profile of degradation resulting species, but also to compare with the QUV accelerated weathering data in terms of similarities and differences. All fonnulations have been kept constant to facilitate comparison.

EXPERIMENTAL SAMPLES Four typical PVC siding capstock formulations with varying levels ofTi0 2 were employed in this study. The formulations are shown in Table 1. The formulations were compounded using

Mapping Studies of Weathered PVC

Table 1. PVC formulations Formulation PVC (K=65) Butyl tin stabilizer Paraffin \vax (MP= 165°C) Calcium stearate Acrylic process aid Acrylic impact modifier Titanium dioxide

100 0.5 1.2 1.5 1.0 5.0

100 0.5

3.0

6.0

10.0

1.2

1.5 1.0 5.0

a high shear mixer. Test sheets were milled and pressed. Typical sample thickness were ofthe order of 60 to 80 mils (1500 to 2000 microns).

OUTDOOR WEATHERING The samples were weathered outdoors in Pennsylvania for a period of 1 year starting January '97. Salnples were removed from the racks at 3 month intervals for analysis.

COLOR READINGS Color reading were performed on a Hunter Colorquest SN 5405 colorimeter. The color of the test chips, after washing, was read immediately after removal from the racks.

FOURIER TRANSFORM INFRARED MICRO SPECTROSCOPY

Mapping Equipment FTIR mapping experiments were acquired on a Nicolet Magna 550 equipped with a Nic-Plan microscope, liquid nitrogen cooled MCT-A detector, motorized micro positioning stage, auto focusing accessory and OMNIC Atlf.ls ™ software.

Sample Preparation Approximately 1 in 2 coupons were cut fronl the test specimens after color reading. The coupon was mounted in a Leitz microtome and 10 micron slices were microtomed from the thickness direction. The slice thickness was adjusted to yield an optimum for the IR band intensities of interest, while minimizing saturation effects. A typical slice was approximately 1500 microns wide by 3000 microns long. Care was taken to minimize damage of the light exposed surface. The slice was mounted on a KBr crystal and placed on the FTIR nlicroscope nlotorized stage.

Weathering of Plastics

FTIR Maps The sample area to be investigated was defined by means ofa set ofknife apertures. The aperture width was set at 18 (± 3) microns. The length ofthe aperture was set at approximately 160 microns. Adjustments in the length were made to maximize signal intensity. A line map was then defined starting from the exposed surface and ending at or near the back of the sample. The map step was a constant 10 microns. The microscope was adjusted to maintain focus throughout the entire experiment. The spectral resolution was set at 8 cm- l and 128 scans were co-added for each spectrum. A typical map consisted of between 120 to 150 spectra. Data Analysis Data analysis was performed using the Nicolet OMNIC Atl~s and Galactic Industries ™ Grams/32 software packages. Standard Methodology The map IR spectra were transformed to absorbance and baselined. Compositional analysis of various species, as a function ofdepth from the exposed surface, was performed using both IR band intensity ratios and band shape fitting. Among the transitions studied were the 1577 cm- l calcium stearate carboxylate, 1780 cm- l acid chloride and the 1718 cm- 1 beta chlorocarboxylic acid. The calcium stearate and acid chloride profiles were obtained using band ratios. The species specific absorbances were normalized to the 2918 cm- l C-H stretching. The change in the beta chlorocarboxylic acid, integrated absorbance band area centered 1 at 1718 cm- , was monitored using curve fitting results. To obtain the integrated area, the c==o stretching region was curve fit using GaussianlLorentzian band profiles.

RESULTS QUALITATIVE DESCRIPTION Figures 1 and 2 illustrate typical results obtained for the PVC formulation containing 0 phr Ti0 2 • The spectra of the unexposed sample exhibit the characteristic absorbances of a PVC compound. Of interest is the c==o stretching region where the acrylic ester band, originating from the impact modifier and process aid, is clearly observed at 1733 cm- 1 and the calcium stearate carboxylate absorbances are noted at 1577 cm- 1 and 1540 em-I. After 3 months exposure, the spectra acquired near the exposed surface have undergone some identifiable changes. The most pronounced are the disappearance of the calcium stearate carboxylate bands and the broadening of the C=O stretching. These changes disappear as one proceeds from the exposed surface towards the back of the specimen. Further exposure, 6, 9 and 12 months results in a progressive loss of the calcium stearate carboxylate absorbance. At 12 months exposure, qualitatively, we do not detect these absorbances (calciun1 stearate

Mapping Studies of Weathered PVC

Figure I. Capslock fonnulation, 0 phr TiOz, no exposure. Line map through the first 100 microns.

Figure 2. Capstock fonnulation, 0 phr TiO z, 6 months exposure. Line map through the first 300 microns.

carboxylate) until approximately 200-300 microns from the exposed surface. Simultaneously, the c=o stretching region broadens on both sides of the 1733 cm- I maximum. On the low wavenumber side a pronounced shoulder is detected at 1718 cm- I , associated with the beta chlorocarboxylic acid and on the high wavenumber side a distinct band appears at 1780 cm- I associated with the acid chloride. For 3, 6 and 10 phr Ti0 2 the spectral changes are less dramatic. We continue to observe the loss of the calcium stearate carboxylate absorbances and formation of chlorocarboxylic acid and acid chloride. These changes are limited to the layers close to the exposed surface and require longer exposure times for detec-

tion. Monitoring double bond formation is not quantitatively possible under the experimental conditions due to low extinction coefficients and overlap with OH and additive associated transitions. Qualitatively, a diffuse set of absorbances do appear during weathering in the characteristic 1600 cm- 1 spectral region indicating fonnation of double bonds. Information regarding polyenes can be obtained from UV spectroscopy measurements. The organotin stabilizer cannot be investigated due to overlapping peaks with the other components of the formulation. CALCIUM STEARATE PROFILE The calcium stearate profiles, measured from the normalized carboxylate absorbance at 1577 cm- I , as a function of depth from the exposed surface is shown in Figure 3 for 0 phr Ti0 2 • As the Ti0 2 1evei in the sample increases significantly less loss is observed as a function ofexposure time. In Table 2 we have attempted to semi-quantitatively determine the depth ofcalcium stearate loss for all samples as a function of exposure time. The values were obtained from

Weathering of Plastics

each individual profile by estimating the depth at which the calcium stearate carboxylate transition returns to its original plateau value. As shown in Table 2, the effect of Ti0 2 level on the disappearance of the calcium stearate is dramatic, varying from deeper than 350 microns at no Ti0 2 to 100 microns at 10 phr Ti0 2 • Table 5 illustrates the previously obtained results under accelerated QUV conditions.

Figure 3. CaSt profile. Capstock fonnu1ation 0 phr Ti0 2, 3, 6, 9, 12 months exposure in PA.

Table 2. Depth (J..lm) of calcium Table 3. Depth (J..lm) of acid chloride forstearate carboxylate absorbance loss mation as a function of Ti0 2 level and as a function of Ti0 2 level and outdoor outdoor exposure time (months) exposure time (months) Exposure time, months

o phr

3 phr

6 phr

10 phr

50 70 100 130

0 80 80 80

0 0 50 50

0 0 0 50

Exposure time, months

o phr

3 phr

6 phr

10 phr

3 6

200 320 350 350

0 100 100 100-125

0 0 50 100

0 0

6 9

100

Table 4. Depth (/lm) of Table 5. Depth (J..lm) of calcium stearate chlorocarboxylic acid formation as a carboxylate loss as a function of Ti0 2 function of Ti0 2 1evei and outdoor ex- level and QUV exposure time (weeks) posure time (months) Exposure time, months

o phr

3 phr

6 phr

120 175 200 250

10 30 70 70

0 0 30 40

o phr

3 phr

6 phr

10 phr

Exposure time, weeks

0 0 0 40

2 4 8 12

200 300 400-450 400-450

50 100 100 100-125

<25 50-75 75-100 100

0 0 50 100

Mapping Studies of Weathered PVC

Table 6. Depth (~m) of acid chloride Table 4. Depth (~m) of chlorocarboformation as a function of Ti0 2 level xylic acid formation as a function of and QUV exposure time (weeks) Ti0 2 level and QUV exposure time (weeks) Exposure time, \veeks

ophr

3 phr

6 phr

10 phr

2 4

0 100 150 150-175

0 25 <50 50-60

0 0 0 0

0 0 0

Exposure tiole, \veeks

ophr

3 phr

6 phr

10 phr

2 4

100 100 300 300

20 40 40 60

0 <20 30 40

0 0 <20 40

ACID CHLORIDE PROFILE (-CH 2 -CIC==O) The acid chloride profile, as a function of depth, was monitored fronl the normalized intensity of the 1780 cm- l absorbance. Formation of the acid chloride oxidation product occurs slower than the loss of calcium stearate. Table 3 illustrates a silnilar analysis to that for calcium stearate shown in Table 2. QUV results are shown in Table 6. CHLOROCARBOXYLIC ACID PROFILE (-CHCI-CH 2 -COOH) The chlorocarboxylic acid profile was analyzed at 1718 cm- I . The results are sUlnmarized in Table 4. QUV accelerated weathering results are illustrated in Table 7.

DISCUSSION The most pronounced effect observed, during the outdoor weathering experiment, is the loss of the calcium stearate carboxylate absorbances. The loss is attributable to the reaction of calcium stearate with Hel to form CaC12 and stearic acid. The loss of calcium stearate occurs deeper into the sample than other detectable degradation processes and is highly dependent on the level ofTi0 2 • Comparing the results obtained under accelerated QUV weathering conditions with outdoor exposure we note qualitatively, QUV exposure results in calcium stearate loss as deep as 450 microns from the exposed surface after 12 weeks for fOl11lulations with 0 phr Ti0 2 • The corresponding 1 year outdoor exposure is linlited to the first 350 microns. Formulations with Ti02, all exhibit calcium stearate loss up to approximately 100 microns at the end of the year, similar to the depth of loss observed for the 12 weeks QUV exposed samples. The behavior of the 1780 cm- I and 1718 cm- l absorbances lags behind the reaction of calcium stearate. A similar effect was observed under accelerated conditions. Both the acid chloride and beta chlorocarboxylic acid are not detectable as deep into the sanl-

Weathering of Plastics

2.40

r-_-:.."TlT~....-_~~~~--;:C:::::.=Q=U=V=1=2=W=ee=k:::::!S =l=;'=;=P=A=12=m=O=~='lh=S:::;JL

~ 1.90...--=--=t~=-.----.--._.- ------t=~-==:: ~

1_40

090

+--------+-----1-7-.....-[-

. . ---.----.. --+-----+--.....- - . { - !

~;:ar:t~.

the

10~he of calc~~~

chlorocarboxylic acid is fonned from the acid chloride in the presence of water, light and oxy-

,:,:= I gen. The behavior observed in

t--t--.----r·-·..-·....-..·.. . . ·-=..~·=t====c!~"'~~..-J

Tables 3 and 4, for the fonnulation containing 0 phr TiO z, PhrTi02 would then have to be explained by the fast reaction, as it is being Figure 4. Comparison of acid chloride and beta carboxylic acid in the first 20 microns for samples exposed to 12 weeks QUV and 12 months outdoor in PA as fonned, of the acid chloride to a function of phr Ti0 2 in the fannulation. the chlorocarboxylic acid. Under accelerated conditions this behavior was observed for all formulations with and without TiOz. In the outdoor experiments the depth profile behavior of the two acid species is very similar. They are detected, at the same exposure time, to approximately the same depth. If the previous behavior were explained in tenns of the fast conversion from acid chloride to beta chlorocarboxylic acid, in the outdoor exposed TiO z fonnulations, the rate of conversion has significantly slowed. As a result the acid chloride species are now detected at 6 phr and 10 phr TiO z where none was observed in the accelerated weathering experiments. Ifwe compare the relative concentration of beta chlorocarboxylic acid and acid chloride in the first 20 microns from the exposed surface, for both accelerated QUV and outdoor exposure we generally find the outdoor exposed samples have higher content. The effect is illustrated in Figure 4. These results would indicate that photooxidation is a more important degradation pathway in outdoor exposure. In comparing QUV accelerated weathering results with outdoor exposure it is important to reiterate the difficulties encountered in making such comparisons. The samples were exposed to non-identical light spectral regions under varying light doses and exposure times. As most weathering reactions induced in PVC are light wavelength dose and time sensitive it is not unexpected that species concentration and distribution will vary. Additionally erosion of the surface plays a crucial role in the analysis as a result of the need to define a zero surface. The data presented assumes no surface erosion and regeneration. 0.40 _ .

·0.10 .L.-.._-:ol-

~3!-=13-.:.-_-__=__=__=_-=-~E.6~f:l:::.======]~;+;J~·£r·=·".".~=<>bl6<:o-~"t-=-..

CONCLUSIONS The results described, clearly demonstrate the capabilities ofFTIR micro spectroscopy mapping techniques to investigate weathering associated degradation processes in PVC fonnulations. Employing these techniques, depth profiles, from the exposed surface, can be obtained for many of the degradation species. Comparing results obtained outdoors with previous

Mapping Studies of Weathered PVC

studies under QUV accelerated conditions we find both similarities and difference. The loss of calcium stearate is the dominant process and occurs to the greatest depth. Fonnation of the acid chloride and beta chlorocarboxylic acid differs depending on exposure type and presence of pigment. Under accelerated conditions the conversion of the acid chloride to the chlorocarboxylic acid appears to occur very fast for all fonnulations. In the outdoor exposed samples this effect is present for the unpigmented formulations. The presence of Ti0 2 decreases the rate of conversion. Additionally, the level of oxidation species, near the exposed surface, is higher for the outdoor exposed sanlples. This may have implications in the deterioration of the mechanical properties associated with weathering. Finally, it is difficult to draw conclusions regarding a direct correspondence between the accelerated and outdoor results in tenns of extrapolating perfonnance expectations.

ACKNOWLEDGMENTS The authors would like to thank D. Tanjala and J. Quiles for the weathering studies, D. Stockwell for sample preparation and R. Tanjala for ICP results. Many thanks to R. Ringwood, and S. Girois for invaluable discussions and suggestions.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

Development in PolYlner Degradation, Vol. 3, ed. N. Grassie. Applied Science, London, 1981. Development in Polymer Stabilization, Vol. 6, ed. G. Scott, Applied Science, London, 1983. Degradation and Stabilization of PVC, ed E. D. Owen, Elsevier Applied Science, London, 1984. (a) J. W. Summers and E. B. Rabinovitch, J. Vinyl Tech., vol 5(3), 91, 1983. (b) E. B. Rabinovitch and R. S. Butler, J. ~1'n)'1 Tech., vol 13 (1), 42, 1991. X. Jouan and J. L. Gardette, PO~l'1n Comm., 28, 329, 1987. J. L. Gardette, Analusis Alagazine, 21(5), M17, 1993. J. L. Gardette, S. Gaumet and J. L. Philippart, J. Appl. Po~)'m. Sci., 48,1885,1993 J. L. Gardette and J. Lemaire, PO~l'm. Deg. and Stab., 34,135,1991. J. Lemaire, J. L. Gardette and 1. Lacoste, Makromol. Chem. Macromol. Symp., 70/71,419,1993. J. L. Gardette and J. Lemaire, J. Vinyl Tech., 15(2),113, 1993. (a) 1. L. Gardette, S. Gaunlet and J. Lemaire, Macromolecules, 22, 2576,1989. (b) S. Gaumet, Ph. D. Thesis, University B. Pascal, France, Decenlber 1989. C. Anton-Prinet, Photochemical Aging of PVC, PhD Thesis, L'Ecole Nationale Superieure D'Arts et Metiers, France, October 1996. D. Garcia and J. Black, SPE VINYL RETEC, Atlanta, GA. 1997, page 183. (a) M. T. Burchill, SPE VINYL RETEC, Atlanta, GA. 1993; (b) R. Ringwood, Private Communication, 1996.

Effects of Water Spray and Irradiance Level on Changes in Copolyester Sheeting with Xenon Arc Exposure

D. R. fa*gerburg and M. E. Donelson East111an Chenlical COlnpan)J, Kingsport, TN, USA

INTRODUCTION We have recently reported the development ofa weatherable copolyester system that involves coextrusion technology of a highly loaded cap layer containing a very effective UV absorber coextruded onto unstabilized copolyester sheeting. 1 In the course of the work to develop this product, we have noted that --e- control .. FL there is a consistent difference between 10 -6-ooot.fol- AZ -'-uv coex,. FL rate of coloring in Florida and Arizona. UV coax ~ A2 But, the difference in coloration rates would appear to be less than the more than 6 35% higher solar flux in Arizona vs. 4 Florida (not to mention the higher ambient exposure temperatures of Arizona) (Fig2 _ ure 1). Haze formation was higher in the Florida exposures (ca. 10% in FL vs. ca. 5% in AZ) in spite of the lower radiation level and ambient temperature. It thus 6 9 12 1'5 18 24 exposure months seemed probable that moisture played an important role in the weathering induced Figure 1. Comparison of b* value in Florida vs. Arizona. changes of our material. The literature on of poly(ethylene photodegradation terephthalate)2-6 shows no provision for any effect of moisture in any of the degradative reac-

81-----" .

/ow·

-:~----: t---r t

Weathering of Plastics

tions although it includes a number of degradation pathways and products. Attempts to sort out effects of radiation intensity and n10isture are best perfom1ed by doing controlled experiments in a xenon arc testing device which is the focus of this effort.

EXPERIMENTAL Coextruded sheeting of our copolyester, poly(ethylene terephthalate) modified with 1,4cyclohexanedimethanol, with a cap layer containing a high loading of an effective UV absorber was performed with a 3.5 in main extruder and a 1.5 in satellite coextruder operated at zone set temperatures of 250 to 260°C. The extrusion was controlled to give an overall sheet thickness of3 Inm (118 mils) with a cap layer thickness of 100-125 microns (4 to 5 n1ils). The takeoff for the extruder was a three-roll stack. Samples of sheeting were cut to 6.35 cm by 14.0 cm (2.5 by 5.5 in) and placed in an Atlas Electric Devices Weather-Ometer, model Ci4000 for intervals of 800 kJ exposure to a total exposure of 4000 kJ. The operating conditions for the irradiance portion of the experiments were: 0.35,0.55 or 0.70 W/m 2 irradiance at 340 run, borosilicate inner and outer lamp filters, 63°C black panel temperature, 550/0 RH with a water spray for 18 min followed by 102 min with no spray. The lamp was not turned off during the spray cycle. For the "dry" portion of the experin1ents the water spray was turned off and a 0.35 W/m 2 irradiance level was used. Color measurements were taken according to CIE recommendations for HunterLab Ultrascan instruments using a D65 light source with a 10° observation angle and specular included mode on the instrulnent. Gel permeation chromatography (GPC) was performed on samples by scraping the exposed and unexposed surface ofthe samples with a sharp scraping tool to a depth of approximately 25 to 50 microns (1 to 2 mils). About 4 mg of each sample was dissolved in methylene chloride-hexafluoroisopropanol azeotrope (ca 70:30 parts by volume) and run on a Perkin Elmer instrument at room temperature through a single PL Gel mixed column. Detection was by UV at 255 nn1. Instrument calibration was established with PET standards of known M w •

RESULTS AND DISCUSSION The irradiance level results showed that the increase in b* with exposure level was lower for the 0.35 W/m2 level than for the other two levels for the copolyester control (Figure 2). For our UV protected coex material, however, the differences in b* with exposure intensity can scarcely be called significant (Figure 3). Haze measurements (Figure 4) showed huge differences between the lowest and highest irradiance levels for both the control and the UV protected material (the intennediate irradiance level, 0.55 W/m2, is not shown for easier reading of the graph but was intem1ediate to the high and low irradiance haze values). In a search for the cause of this difference, we measured the GPC of the exposed surface. The loss in M n vs.

Effects of Water Spray and Irradiance Level

-e- 0.70 W/rn21 -it- 0.55 -a.-

W/m2

0,35 Wfm2

I 1.5

O¥---~-----'''''-''''''--..,...-----r--~

800

2400

1600

3200

1600

4000

exposure. kJ

Figure 2. b* value vs. irradiance level (control sample).

80-.r---....L----"""'---..L-----.J----4 70

-a-O,70 WJm2· con1rol -'-0.35 Wfm2 . control .......... 0.70 W/m2 - UV coex " 0,35 V.//m2 • UV

2400

3200

4000

exposure kJ

Figure 3. b* value vs. irradiance (UV- protected sample).

1.610 4-t-_ _""--_ _....t-_ _-l...-_ _--i....._ _-+ 1.4 10

coax

40 8000

6000

4000 2000"t-----r-'""--~- .............,.----3..,..·O-O--40..00 exposure kJ l

Figure 4. Haze vs. irradiance.

Figure 5. M n vs. irradiance (control sample).

total exposure (Figure 5) did not appear to be a function of the irradiance level. The branching, as measured by the Mz/M n (Figure 6) was difficult to interpret very well but did not appear to support significant differences in the branching behavior vs. the level of irradiance. Although the color data and lTIolecular weight data appeared most probably to obey reciprocity, the haze definitely does not obey reciprocity when the irradiance was increased. It is

Weathering of Plastics

-e-contrc·J " \*let 14

-control .. dry -'-UV protected . . wet . UV prot~<;ted dry y

800

1600

2'00

B}l',pOSUro,

Figure 6. Mz/M n vs. irradiance (control sample).

k.J

3200

4000

Figure 7. "Wet" vs. "dry" conditions. b* value.

well to note in this connection that results obtained by us at the 0.35 W/m2 irradiance level seem to fit Florida outdoor data for b* value but that the haze in Florida exposure is more on the order of 10% rather than the 30% suggested by these experiments. The Arizona data is not appreciably higher in b* which fits with the color data at higher irradiances. Haze in Arizona is only about 5%, however, which does not fit with this data as higher irradiance in the Weather-Ometer caused even higher haze. Thus there must be another explanation for this observation and further work is needed to understand the haze effects on exposure. This brings us to the next item ofconcern, the effect ofwater during exposure on our materials. By eliminating the water spray on the samples ("dry" conditions), there should be effectively a much lower specimen humidity at the surface. Surprisingly, the b* value under these exposure conditions showed (Figure 7) a much higher rate of coloration than water sprayed samples ("wet" conditions). This was true whether the material was the unprotected control or the UV-protected coex sheeting. The haze during exposure was higher "wet" vs. "dry" for the control sample but the opposite for the UV-protected sheeting. This is not currently understood at all. This was also not consistent with the outdoor data which implies that more humid environments give more haze for either unstabilized or stabilized samples. Thus, some other factor is at work in the outdoor weathering results other than purely humidity effects. GPC showed that the "wet" condition samples had consistently higher loss in Mn with equal exposure compared to the "dry" condition samples (Figure 8). The branching, measured by Mz/M n was essentially the same for both sets of conditions. A possible explanation

Effects of Water Spray and Irradiance Level

35t;::========~_-1-_--'-----t

1.610 .;-----''-----'-----'----'----t

-e- control· dry 30 25

~ control· wet .........UV protected· dry UV protected· wet

1.410~" 1.210'

20 15

8000 Mn

6000

-e-wet . front ....wet· back -dry' front dry· back

haze, % 5

4000

o¥o=--""'80~0--1"":'60""'O""-~2""40"""O--""'32""'O-0--4-+000

200°:1--....".,,,,,---=,.,,--=,,,---3=00::-----04"*000

exposure, kJ

Figure 8. "Wet" vs. "dry" condition haze.

Figure 9. "Wet" vs. "dry" condition Mn values (control).

of these observations is that there was photohydrolysis occuring. It is reported 7 that aryl terephthalates undergo photohydrolysis but not that the alkyl esters do. The light source in the literature experiments was, however, a mercury arc with a spectral power distribution that is far different from that ofaxenon arc source. Thus, in our experiments, the possibility ofexcitation of the carbonyl chromophore which would then possibly lead to photohydrolysis was certainly a possibility. This then argues for a mechanism during photodegradation where a photon is absorbed in the aromatic compound to produce an excited state. This excited state can then do multiple things, of course, but one path would involve production of a colored species and another competing path would involve reaction with water to give photohydrolysis. If there is little water around, then the predominate path of reaction would be color formation. If, however, there is a much higher concentration of water present, as in the "wet" condition, then photohydrolysis competes effectively for the excited state. The products from the photohydrolysis are, of course, colorless.

CONCLUSIONS Reciprocity is obeyed reasonably well for the color (b* value) and also the molecular weight data for unprotected and also UV-protected copolyester sheeting. Haze, however, failed to obey reciprocity and additionally gave less than consistent results between the two sets of samples tested. There would seem to be an additional mechanism operating that causes haze

Weathering of Plastics

in Florida and not Arizona. Exposures in the absence of water spray raise the possibility of photohydrolysis in this systen1 in competition with the reaction(s) responsible for color formation.

ACKNOWLEDGMENTS The authors would like to acknowledge the able assistance of many coworkers, among whom are Helen Richards, Brenda Forbes, Edgar Gamble, Joda Wood, and George Caflisch.

REFERENCES 1. 2. 3. 4. 5. 6.

D. R. fa*gerburg and M. E. Donelson, J. Vinyl and Additive Tee/mol., 3, 179 (1997). J. G. Pacifici and 1. M. Straley, Polym. Lett., 7,7 (1969). G. Valk, M.-L. Kehren, 1. Daamen, Angelt: AIakrom. Chemie, 13, 97 (1970). M. Day and D. M. Wiles, J. Appl. Polym. Sci., 16,203 (1972). P. Blais, M. Day and D. M. Wiles, J. Appl. Polym. Sci., 17,1895 (1973). For a synopsis see for example "Polymer Photodegradation, Mechanisms and Experimental Methods", Jan F. Rabek, Chapman and Hall, London, p288-94 and references cited therein.

Hot Water Resistance of Glass Fiber Reinforced Thermoplastics

Takafumi Kawaguchi, Hiroyuki Nishimura, and Fumiaki Miwa Osaka Gas Co., Ltd., Japan Kazunori Ito, Takashi Kuriyama, and Ikuo Narisawa Yanlagata University, Japan

INTRODUCTION Recently, the casing of parts in appliances for hot water supply such as valves, pumps, and sensors are made ofplastics instead of metals for the purpose ofreducing cost and making appliances lighter. Glass fiber-reinforced plastics (GFRP) are nonnally used for those parts because their internal pressure is sometimes high. Such parts are also required to be resistant to hot water for a long time because these parts are always exposed to hot water. In Japan, glass fiber-reinforced polyphenyleneether (GF-PPE), polyphenylenesulfide (GF-PPS), and polyoxymethylene (GF-POM) are nlainly used for such parts. There are SOlne reports on the water resistance of GFRP or carbon fiber (CF)-reinforced plastics. I -3 These reports were focused on OF-reinforced epoxy. 1,2 CF-reinforced epoxy, and polyetheretherketone,3 and OF-reinforced vinylester. 4 In this study, the authors investigated the hot water resistance of these glass fiber-reinforced thermoplastics by measuring the change in tensile strength when they were immersed in hot water. OF-PPS was also exposed to hot air and the change in tensile strength was measured. For comparison, hot water immersion tests were also performed on materials which were not reinforced. The surfaces and the tensile fracture surfaces of the specimens were observed by a scanning electron microscope (SEM) and the cause of the change in tensile strength was investigated.

Weathering of Plastics

100

EXPERIMENTAL MATERIALS AND TEST SPECIMENS

Table 1. Materials used in this study Matrix resin Polyphenyleneether Polyphenylenesulfide Polyoxymethylene

Glass fiber or beads content, 0/0 20 (fiber) 30 (fiber) 25 (beads)

The materials used in this study are shown in Table 1. The surfaces of the glass fibers and the glass beads contained in the materials were treated by silane coupling agents. The test specimens used in this study were ASTM D638 Type I and were injection-molded. The thickness of the specimens was 1.60 mm.

HOT WATER IMMERSION TESTS The specimens were immersed in hot water maintained at 90°C up to 9000 hrs. The change in weight was also measured during the tests. These specinlens were used for tensile tests and SEM observations. For comparison, GF-PPS was exposed to hot air maintained at 90°C and the change in tensile strength was also measured.

TENSILE TESTS For the tensile tests, an Instron universal testing machine type 5567 was used. The tensile speed was 50 mm/min. Tests were perfonned at 23°C and 50%RH, and the specimens were kept in the same condition for 24 hrs before the tests.

SEM For SEM observations, a Hitachi SEM S-530 was used at an acceleration voltage of 20 kV. For observing their surfaces, the specimens for SEM observations were cut out from the specimens after they were used for the hot water immersion tests. The specimens used for the tensile tests were also used later to observe their fracture surfaces.

RESULTS AND DISCUSSION MEASUREMENT OF CHANGE IN TENSILE STRENGTH Figure 1 shows the change in the tensile strength of the GFRP specimens. It was found that the strength of the specimens decreased when they were immersed in hot water. The tensile strength ofGF-PPS showed a remarkable change and its strength after 9000 hrs of hot water immersion was only about 57% of its initial strength. The tensile strengths of OF-PPE and

101

Hot Water Resistance 80

160 -'-GF-PPE

140

--.- GF- PPS .... GF·-POl-l

ltl 120 0.

;;;

tc: fj

ltl 60 0.

S~1

o 40 1Il

I))

_" -" - _- __.- ..- ..- .

;;; 50

100

S1Il

r-----

o o

2000

4000

6000

8000

-o-PPE

.. n·PPS

-t:r- PO!1

1...-

10000

•

2000

4000

Tine Ihn

6000

8000

10000

Tine Cn;)

Figure I. The change in tensile strength of specimens of Figure 2. The change in tensile strength of specimens ofGFRP. GFRP.

lBO

160 ~ ()~

140

.. , ,

-GF,PPS;

120

-50'

e 100

g III

...,

:2,

III

,.,c: E--

:?;';

- G F PPS

40 20

C7.L G (1

[)

'--

0 0

2000

4000

6000

8000

10000

T~Q~~

Figure 3. The change of tensile strength of GF·PPS when exposed to 90°C air.

1000

2(lOO

3000

4(}()()

T~e~H

Figure 4. The variation in weight of GF·PPS during the hot water immersion test.

glass bead-reinforced POM (GB-POM) after 9000 hrs of hot water immersion was 88% and 87% of their initial strengths, respectively. Figure 2 shows the change in the tensile strength ofspecimens which were molded out of non-reinforced plastics when they were immersed in hot water. Although the initial tensile

102

Weathering of Plastics

Figure 5. The SEM images of the surface ofGF-PPE (a) before the hot water immersion test and (b) after 9000 hrs of the hot water immersion test.

strength was small compared with that ofGFRP, its tensile strength did not decrease. Figure 3 shows the change in the tensile strength of GF-PPS when it was exposed to 90°C air. It showed only a little change in tensile strength. Figure 4 shows the change in the weight ofGF-PPS during the hot water immersion test. It was found from the results that the change in the weight ofGF-PPS during the hot water immersion test was small. SEM OBSERVATIONS

Figures Sa and Sb show the SEM images of the surfaces of glass fiber-reinforced PPE (GF-PPE) before the hot water immersion test (Figure Sa) and after 9000 hrs of the hot water immersion test (Figure Sb). It was found that the surface of the GF-PPE specimens after the hot water immersion test was rather rough compared with that before the test. In Figure Sb, debonding between the glass fiber and the PPE matrix and surface cracks are observed. The cause ofthese debonding and surface cracks seemed to be the residual stress on the surface. On the surfaces of the GF-PPE specimens after the hot water immersion tests, flow marks could be recognized clearer than before the tests. The appearance of the flow marks seemed to be attributable to the debonding between the glass fiber and the matrix resin, which resulted in the clear appearance of the glass fiber on the surface. Figures 6a and 6b show the SEM images of the tensile fracture surfaces ofGF-PPS before the hot water immersion test (Figure 6a) and after 9000 hrs of the hot water immersion tests (Figure 6b). In Figure 6a, the bonding between the glass fiber and the PPS matrix was rather good. On the other hand, in Figure 6b, it was found that only a little matrix resin re-

Hot Water Resistance

103

Figure 6. The SEM images of the surface ofGF-PPS (a) before the hot water immersion test and (b) after 9000 hrs of the hot water immersion test.

mained on the surface of the glass fiber and that the bonding between the glass fiber and the matrix resin was rather poor. Figure 7 shows the SEM image of the tensile fracture surface of GF-PPS which was exposed to 90°C air for 6500 hrs. The specimens had nearly the same strength as the initial strength after the tensile test. In Figure 7, the bonding between the glass fiber and the matrix resin was rather good. DISCUSSION ON THE CHANGE OF TENSILE STRENGTH Figure 7. The SEM image of the fracture surface of GF-PPS h h 'd . I h after exposure to 90°C air for 6500 hrs. T e aut ors pal specla attention to t e change in the tensile strength ofGF-PPS, because its change in tensile strength was most remarkable in the hot water immersion test (Figure I). From the SEM observation of the tensile fracture surface, it was found that the bonding between the glass fiber and the matrix PPS resin became poor by hot water immersion. The authors concluded that the change in the tensile strength ofGF-PPS was attributable to the deterioration ofthe interface between the glass fiber and the matrix PPS resin. The conclusion was also supported by the fact that the tensile

Weathering of Plastics

104

strength ofthe PPS specimens, which were not reinforced, was very small compared with that of reinforced ones. It was also found that the change in the weight of the GF-PPS specimens was very small during the hot water immersion test. It was concluded from the results that the deterioration of the interface between the glass fiber and the matrix PPS resin was caused by a small amount of vaporized water which penetrated into the specimens.

CONCLUSIONS The hot water resistance of three kinds of short glass fiber or glass bead-reinforced plastics (PPE PPS, and POM) was studied by hot water immersion tests and tensile tests. It was found that the tensile strengths of these plastics all decreased and the change was most remarkable in GF-PPS. SEM observations ofthe tensile fracture surfaces revealed that the change in tensile strength was attributable to the deterioration of the interface between the glass fiber and the matrix resin. The cause of the change in tensile strength was concluded to be the penetration of vaporized water into the specimens. Although the change in the tensile strength of OF-PPE was small compared with that ofOFPPS, the debonding between the glass fiber and the matrix resin PPE and cracks were observed on the surface of the GF-PPE specimens.

REFERENCES 1. 2. 3. 4.

Carl R. Schultheisz et aI., Composite Materials, 1997,113,257. L. Salmon et aI., Composites Sci. and Techno!., 1997,157,1119. R. Selzer and K. Friedrich, Composites Part A, 1997, 28A, 595. S. Sridharan et aI., ANTEC'98, 2255.

Surface Temperatures and Temperature Measurement Techniques on the Level of Exposed Samples During IrradiationlWeathering in Equipment

Jorg Boxhammer Atlas Material Testing Solutions, Gernlany

INTRODUCTION During both natural and simulated irradiation/weathering, the surfaces of exposed material samples become heated above the temperature of the surrounding air. Nearly all photo-induced aging processes are influenced by the surface telnperature of the salnples. For this reason, the surface tenlperature represents an extremely important parameter for the itTadiation and weathering of polymer materials. Direct measurement of the surface temperatures of individual samples is not practical in weathering equipment. Therefore, one method that has been enlployed for quite some time during simulated weathering involves the measurement of the temperature of a "blackened, flat panel", described as the "black panel temperature or BPT". There are various designs and models of these types ofinsttuments (employed in different types of weathering equipment), exhibiting systemic differences and, to some extent, providing inadequate degrees of measurement accuracy. Thus, the goal of optimizing the reproducibility of test results in weathering equipment cannot be achieved with these measuring elements. In addition, it has also been found that the desired characterization ofthe highest possible surface temperature - in other words, the maxinlum absorption by solid samples - cannot be realized with these measuring elements. The above considerations formed the starting point for the development of a new measuring element, known as the "black standard thermometer or BST" in order to differentiate it from the former elements. This new element is not subject to the disadvantages inherent in the previous one. Both the theory and the technical design of this measuring device are under disCUSSIon.

Weathering of Plastics

106

Currently, both black panel as well as black standard thermometers continue to form the basis for the temperature requiren1ents contained in numerous national and international standards and are correspondingly utilized in weathering equipment. While current standards generally note that the various measuring elements will produce different temperature readings under identical conditions ofexposure, systematic relationships and measured data have, thus far, been inadequately published. For this reason, currently available measured data are included. The BST is necessary, but by no means adequate, for the characterization of the surface temperature ofexposed materials. The additional measurement ofa so-called "white standard temperature" is also required, and is currently under discussion. In practice, during simulated weathering, the sample chamber temperature is measured and held constant to act as a substitute for the white standard temperature. In modem weathering equipment, the measurement and regulation of both the BST and PRT by means of a variable fan speed represents an assured method ofexamining unchanging temperature levels, even under conditions of varying UV irradiation strength. The technical relationships will be explained.

SOLAR RADIATION AND SURFACE TEMPERATURE OF POLYMER MATERIALS 10000

10tXJ

1 '00

1fJ

'0,' Figure 1. Spectral irradiance vs. wavelength (VDI 3789, Blatt 2). Curve 1. Solar radiation (short wavelength radiation). Curve 2. Terrestrial radiation of black body having telnperature of 300K -27°C (long wavelength radiation).

The spectral irradiation strength of solar radiation on the earth surface depends on both the time of day as well as the season. An understanding of material aging in practical applications or during natural weathering, as well as the standardization or methodical planning of tests under simulated conditions, require as precise as possible a level of knowledge of the natural radiation conditions. Environmental meteorologyl differentiates between solar radiation and terrestrial radiation. Radiation from the sun corresponds to the radiation from a black body with a temperature of roughly 6000K, while the heat generated by terrestrial bodies corresponds to those with a temperature of roughly 300K. Since the associated spectral ranges (refer to Figure 1) are almost totally separate from one another, solar and terrestrial radiation are usually also treated separately, and are respectively designated as short-wave and long-wave radiations. With respect to material testing, the solar radiation reaching the earth surface is of primary im-

Temperature Measurement Techniques

107

portance. However, the thennal radiation must also be taken into account with respect to the temperature on the Inaterial surface. In order to be able to assess the silTIulation quality of artificial radiation sources or radiation systems, a reference spectIum must be defined. For today's test purposes, specific data concerning the spectral distribution of the "global radiation" (the SUln of direct and diffuse radiation) on the earth surface are ~ ~ aoo W~::l&fl~~nm~fM ,~ 2m WA employed (CIE No. 85:1989; Table 4 2). Figure 2 illustrates the spectral irradiation Figure 2. Spectral irradiance (erE No. 85: 1989; Table 4 2). strength. In contrast to the field of meteorology, material testing concerns itself with only the short-wave portion of global radiation (UV + a portion of the visible range), while the long-wave component of global radiation is represented by the IR range. Basically, only that portion of the absorbed radiation corresponding to the material's specific absorption level, E(A), (generally dependent on the wavelength) will produce an effect. A very small portion of the absorbed radiation will produce prilnary photochemical processes. In this regard, the UV range of solar radiation is of primary significance. Due to the complex nature of the wide variety of material systems and the diversity of wavelength ranges they absorb, as well as the fact that other and different spectral portions may also be responsible for the alteration of various material properties, it is generally recommended, ifnot required, that the simulated radiation be adjusted to reproduce as wide a range of solar radiation as possible. The n1ajority of the radiation absorbed by a material results in the material's temperature being raised above that of the an1bient environment. The temperatures of dark sample surfaces can be as much as ~ T=60K higher than the temperature of the surrounding air. This temperature increase depends primarily on the irradiation strength, the thickness, and the optical and thermal characteristics of the sample material, as well as the ambient climatic conditions (wind velocity/humidity) at the irradiated surface. The radiation exchange with the environment must also be taken into consideration. Measurements of variously dyed PVC foils in the open and set at 45° angles facing south - both with and without a backing (insulation) - resulted in temperature differences of 15 to 23K between black and white. 3 Figure 3 illustrates an example. In an ideal, non-transparent material, the radiation is absorbed only at the surface, where it is converted to heat. In this case, a linear temperature gradient establishes itself between the higher temperature level on the surface and the unexposed back portion of the sample. Con200

1000

2000

Weathering of Plastics

108

tCR£ElI ~

ORMIG[

, llUOW

OR(D ~R[r[p.wC[

WKI1£

0 -1 ~

.5 I-

a..

m $3 Ql

~ 'C

....e ;:)

C1l

-2 -3

2Dl Q

•

6B 85 102 119 135 is]

1i'~

---........ rr--'-'... ,;..

~~(

-4

c-=:

-6

.--·7 .__. " Blue - -

,..-

-9

::;-Red

·10

-CXange

·11

·12

r-f--

r----

-6

Green 1 -

--lZI Yellow •...

f--

-5

Q. ·13

r±~".--<... "'-"

LL ".y..

f--

r-f--

'7'l" t== ._.

J..

r-f-f--

:Q:

OWlne ~

.....

·14

.15

-16 - ' - - - - - - - - - - - - - - '

MINUf£S

Figure 3. Temperatures of PVC films of different colors.

trasting to this are cutaway samples and/or those with pronounced surface structures where the maximum temperature arises not at the sample surface, but instead (depending on the material's absorption coefficient and its surface structure) at deeper layers in the materiaL In such cases, there may be a cooling effect produced by air flowing over the sample since heat transfer is primarily a function of thermal conductivity. In such materials, significantly higher temperatures may arise below the surface than is the case with solid samples with absorbent exterior surfaces. Nearly all secondary decomposition processes of organic materials that occur as a result ofthe primary photochemical step are temperature dependent. In general, the speed ofthe decomposition reaction increases with increasing temperature. The approximation rule from the field ofreaction kinetics states that a temperature increase of 100 e will lead to an approximate doubling of the reaction speed. This has been confirmed for, for example, photo-oxidation of polyolefines. 4 In many cases, no linear relationship is observed, but instead, a material- specific and increasing degree of alteration with increasing temperature. Figure 4 illustrates an example of this - the color change in an irradiated pve sample at different irradiation levels and as a function of temperature. The actual surface temperatures of the samples are not directly measured, but are instead characterized by the black standard and white standard temperature for each temperature level (sample chamber temperature) (refer also to Section below).

Temperature Measurement Techniques

109

4r---~--,-----,---r---,

i -+- H 1 3,5

2000KJlm" :-+- H 3000 KJ/m' : , l-.- H 5000 KJlm' , i , __ H 8000 KJ/m' i;"

2,5

4l!

: -fI- H 10000

KJlm':

95 .r=-,-_,_---,--......-, 90

,~H

000 ~·J·/m1 :[;i

" 75 ~ 70 ;;, 65

~ ~~

',.,'_.,-

I··"",K

50 45 I·· 'j·,··;-·i, 40 I", , , .. ; ;

__ H 15000 KJlm'

1.5

,--,---j

!---''-----'---'-'--l.-J 35 40 45 50 55 60 65 70 Chambtr temperature In 6C

0,5

Xenotest Alpha LM; FS XOOO

01-----'------'----'---4--'-----1 ~

CT in·C

E (300-400nm) - 100 W/m'

ReI. humidIty

20%

LDZ = 2000 l/min

Figure 4. Effect of temperature on photochemical degradation of PVC at different irradiation levels and as a function of temperature.

The second order transition or glass temperature, T g, also plays an important role because the mobility ofthe chains and radicals increases above the T g while, simultaneously, the oxygen diffusion coefficient changes, The effect on the decomposition speed depends on the active ingredient and the specific decomposition process in question. 5 Based on the relationships indicated above, the temperature of the irradiated surface at which most of the decomposition processes occur is a far more important parameter than the temperature of the surrounding air for the irradiation and weathering of polymer materials. It is therefore vitally important that: • The surface temperatures of exposed materials be carefully characterized during simulated, time-lapse illumination/weathering in equipment, and; • The surface temperatures resulting from the effects of solar radiation - which are material-specific and can vary widely - be adequately adjusted during the simulation. This can only be achieved by means of a radiation continuum that adequately simulates both the predefined requirements for the photochemically effective spectral range as well as the total spectrum and level of the radiation.

TEMPERATURE MEASUREMENT AT SAMPLE LEVEL IN WEATHERING EQUIPMENT In time-lapse tests in equipment, the thermal effect of the radiation is most practically measured in conjunction with the other climatic parameters, and is given as the "surface temperature at the exposure level".

110

Weathering of Plastics

For special tests, e.g., in conjunction with the determination of kinetic reaction data, direct measuren1ent of the surface temperatures of the exposed samples is indispensable. In doing this, the following possible sources of measuren1ent errors must be carefully considered: • In optical measurements, radiation reaching the sample surface as a result ofreflection, or; • Where adhering or embedded temperature sensors are employed, the absorption of the sensors themselves as well as the thermal conductivity of the sensor wiring. In general during illumination/weathering tests in equipment, direct measurement of the surface temperatures of individual samples is not practical. Instead, limit conditions are described, based on the observation that, where all other environmental conditions are equal, the highest degree of heating will be exhibited by a material with an absorption level of E == 1, while the lowest degree of heating will be exhibited by one with an absorption level OfE == O. Based on this, the thermal state at the sample level or the range of surface temperatures of samples under given environmental conditions can be designated by appropriate measuring elements. 6 A common practice in simulated weathering has been and remains the measurement and regulation of the temperature of a flat, black panel. This is known as the "black panel temperature or (BPT)". This method can, for example, easily detect the utilization of an incorrect filter systenl. However, in the past, based on the realized prerequisites for nlaxilTIUm radiation absorption, this temperature was often incorrectly equated with the "maximum surface temperature of exposed samples" under certain climatic conditions. As experience has, however, shown, the actual surface temperatures of solid, black or dark samples can be significantly higher than the BPT. This is linked to the danger of undesirable or unrecognized aging processes. In addition, side-by-side comparative measurements in a single weathering device between the various, structurally differing, black field thermometers (the basic design is illustrated in Figure 5) employed in different weathering equipment have shown that significant differences in the displayed tenlperatures can arise, particularly as the irradiation strength increases and at higher temperature levels. 7 These effects are documented in the numerous measurements performed in conjunction with the revision or preparation of standards by the various DIN working groups in the area ofillumination/weathering ofplastics and elastomers or for materials designed for automobile interiors 8- 1o as well as in the not yet completed comparison tests in ISO/TC61/SC61 WG2. 11 The consequence arising from this is that, under formally identical BPT adjustment conditions in different types of weathering equipment, the actual surface temperatures of exposed materials can differ significantly, resulting in variations in the material and property-dependent test results.

111

Temperature Measurement Techniques

BidtwtMg-.iir1f-ffi

... ,~P.J.ld .....

"'~If!:t;rtlj.iC!)

Jt!~h~I$i<J1.¥ijl

. ~

,/;/ ~~ /

I1r

: ci'o.Wm

d;j¥.:c )(llW},n1)7,(r.!ft)

Figures 5 and 6. Schematic diagratns of black panel thennon1eter (left) and black standard thennometer (right).

IMPROVED TEMPERATURE MEASUREMENT SYSTEM • BLACK

STANDARD THERMOMETER The above problelTIs formed the starting point for the development of an improved measurement principle that, in order to differentiate it from the older elements, is known as the "black standard thermometer or BST" and, with regard to its basic structure, is based on a recommendation of the BAM (Gemlan Federal Institute for Material Research and Testing). For a more detailed explanation of the theoretical and experimental foundations, please refer to the relevant literature. 6 The following seeks only to summarize the most important factors. The primary goal in developing this new procedure to develop a concept for measuring the temperature ofa nearly dimensionless, thin surface layer with high thermal conductivity for measurenlents of the surface temperature of exposed, highly absorbent samples. The ideal, dimensionless surface can be adequately simulated by a black coating of minimum thickness and a high degree ofthermal conductivity. Ifthis coating is applied with a thickness of, for example, 5 mm, to a polymer sample with low thermal conductivity, the ov~rall energy balance remains generally unaltered. Only the value ofthe heat flux density from the surface to the interior of the sample - approximately 30% of the heat flux density of the arriving solar radiation - becomes inconsequentially smaller. As a corrective factor, only the small, additional thermal capacity of the "new" surface must be taken into consideration. This value delays the temperature increase over time, but the lTIaximum temperature remains unchanged for a constant in adiation strength of short-wave solar radiation. 4

112

Weathering of Plastics

Figure 6 contains a schematic side view ofthe design ofa new measuring element based on the above model. The ideal. "dimensionless" surface of a 5 mm thick sample exposed to the radiation is replaced by, for example, an approx. 0.5 mm thick panel made of stainless steel with a black coating. In order to minimize margin effects, the panel must have a minimum diameter of40 mm. For practical reasons, a panel with a 40 mm x 70 mm surface area is specified. A preferably electrical temperature sensor is mounted on the rear ofthe plate in a manner that ensures good thermal conductivity. The sensor wiring is also installed so as to offer good thermal conduction, and runs along the back of the panel to the panel edge. The metal panel is mounted on a 5 mm thick base panel made ofPVDF and exhibiting an indentation in the area of the sensor element in order to avoid direct contact between the sensor and the plastic panel. This base panel simulates the thermal insulation produced by a solid polymer sample. This design results in a "clear", black surface with a sensor protected against direct radiation whose input is heated by the black panel thus minimizing conductance errors. Because of the metal panel's good thermal conductivity, this design only differs from that of an ideal, "dimensionless" surface in its thermal capacity. For this reason, there is a slight delay in the temperature increase after the radiation source is turned on. The black paint, whose spectral reflection parameters are specified, 7 is in tum covered by a coating of clear lacquer which does not alter the absorption capability but does improve the panel's resistance and allows the panel's functionality to be fully restored after weathering by polishing. Experience gained over several years has shown that the suitability of the surface coating is limited during continuous use (measurement and regulation of the BST), particularly under extreme conditions (i.e., as specified in DIN 75202). A black chrome layer such as is employed in the Xenosensiv measurement system (a combined system for UV irradiation strength and BST, manufactured by Atlas) has proven itself in practical applications. The system design of the BST in various weathering equipment models available is basically identical, even thought the overall design and layout of the equipment varies greatly. Figure 7 illustrates some examples. For continuous measurement and regulation of the BST in carousel devices, the signal from the measurement instrument located on the sample level is transferred via loop rings. Figure 7. Examples of black standard thermometers.

113

Temperature Measurement Techniques

STANDARDIZATION OF TEMPERATURE MEASUREMENT DEVICES There is no specific standard for the black standard thermometer. However, based on a description of its primaty technical components, the measuring element was included in major DIN standards (DIN 53 387; DIN 53 231; DIN 75202) by the end of the 1980's and, subsequently, in the conesponding ISO standards (i.e., ISO 8492; ISO 11341). The latter however also still permits the use of black panel thermometers (the type employed must be identified in the test report). Some specifications whose temperature requirements are based solely on the BST also still exist. In general, the goal of improved reproducibility of the test conditions in weathering equipment can only be achieved by measurement instruments that display identical values under a set of given exposure conditions. However, since this goal cannot currently be met, a detailed understanding ofthe display differences between various BPTs and, in particular, between the BPT and the BST, as dependent on the weathering conditions, would appear to represent a minimum requirement. This knowledge can then be taken into account when conditions are set "to the same level" as well as during comparison and interpretation of test results from different equipment.

COMPARATIVE DATA FOR TEMPERATURE MEASUREMENT INSTRUMENTS Essential data was prepared during the course of comparative tests in ISO/TC61/SC6/WG2Light exposure; a test series initiated in 1994, and cunently still underway. 11 During an initial phase, comparative measurements 1~1 lOll F'"ir~~"'P;--:t:;:'::::;:~9 were made with commer91 9Ii f-----tf--+7=t__ cially available BPTs and 9-! 91 BSTs in various weathering 90 t--T-h'--t---j 88 I--i-)'o----+-+----t----jequipment and under system1-' 66 ~_f/_7_--+-+--t---fatically varying conditions. ~ ~ t.··J;:.~:.:::j=±=±=t! Figure 8 illustrates the ~ ~ R dynamic behavior of two ;~ - calculated (exp fun ) :/ BSTs from different manun70 !:__ . _-II facturers for a jump function 1000 1200 o 100 400 600 800 1000 1200 in inadiation strength at a Time i1sec. Time i1sec. continuously regulated sam- - - ' L.. •••.•_ _•.....•..•__

m. .,=••., ~,

_c==:::===-_~._

Figure 8. Dynamic behavior of two different black standard thermometers (jump function E (UV) = 45 at 120 W/m 2, CT = 50°C).

pie chamber temperature. The curves can be adequately

114

Weathering of Plastics

described by an exponential func94 tion with a 1 92 1 negligible differ911 'i--+-h~'+-"""'~"'t'\i1I 88f--j----it''--+--t---t----i ence. Figure 9 113 ---.-- ---contains corre~64l--·- ~82 t----\I'--t---+--t----1r-----1 sponding curves Ii:: I'IJ f--It--f--+--+----1r--I for two BPTs 10 78 (electrical signal 76 74 acquisition). Surn prisingly, the 70/·····1···············,······ temperature ino 100 200 300 400 500 600 crease with which Tlmeinsec. the BST is compaFigure 9. Dynamic behavior of two black panel thennometers. rable - although displaced in the direction of less time required to reach 105 •. Xenotest 1200 LM Filtersyslem 3 Supra>< dishes equilibrium - lies at a distinctly lower 100 f- Chamber temp. 50'C temperature level and differing display 95 I- Ret humidity 20 % 90 f- Frequency 50 Hz on the BPT. I-' 85 I- Semples stainless steel .-.< ""~' In-I ...-~;.:,:,~~ _;r;f~~4 ! The difference between BPT and .E
SIll

---

115

Temperature Measurement Techniques

xenolesl Bela LM

LDZ= 1500 llmln

.E 75 70--__""S-"'T-c;7·-'-':-:···i,:c,·I·~:O:···· ....················

!................

.L-_ _----\.

-+-e1ektr.;MWA.S --manual; MWA-S manual, Xl -+------J,

:m.:.:,.~a.:n.:.~u:.~a.,I::;.X2""

() 85 o

E(UV) and PRT vaned

.5 Ii;

85 I . +

tf/

80 I

I·······.. :

__.j,

;

II X 45°

"f"

65 65

100

A in·C

Figure 12, Three BST under varied conditions of exposure.

r--_,_-,---,-----,-~----,---,--.,....,

Xenotest Beta LM FS XC300 ReI.H." 20% E(UV) and PRT vaned

80· --·---r-·-·.. 75

-' _..[!• s I

+4=~

.. !

,.........!

SST

Figure 11. Different BPTs .. temperature as a function of E(UV),

-.;..--

70 .J.J

E(300-400nm) in W/m 2

..f

120

100

FSXC300 ..: RelH; 2 0 % 1 ; )

III

65 ..•"" 60 I....··•• .. ······.. ·· 55

.... Xenolesl Bela LM

100

FSXC300 PRT= 50'C RelH.a 20%

()80

105

.-----~-------,----_,_---"

.- .... --.....

•

I >i,'~' b~'

1 ',.J":;'>!

.~ >1';'" I ,~~i__ li~:;~~;,:

instruments (S and X). The deviations across the entire range ofvarying conditions are systematically very small. The same ap-

plies correspondingly for the BPTs with electrical signal acquisition, where the difal 45 65 ference is, however, somewhat larger. This 60 />+' . r; 0.9942 is also true for BPTs with manual readout of 55 . , ' , : the same type, but with large differences (up 55 60 65 70 75 60 65 90 95 to 9K) for BPTs ofdiffering design (data not BPT el. A In °C shown). In view of such BPT temperature difFigure 13. Three BPT under varied conditions of exposure, ferences and the indicated differences between BSTs and BPTs, it is apparent that, for these differing measurement systems, "the same indicated temperature" cannot be equated with "identical surface temperature" of exposed (black) samples, which is one of the essential prerequisites for reproducible adjustment of conditions in devices. Gi

/,', ,,.'1 ! " " . i ,~' ,.~' I

SURFACE TEMPERATURES OF EXPOSED MATERIAL SYSTEMS While the adjustment of identical black standard temperatures is a prerequisite for the creation ofadequately similar surface temperatures on exposed samples under natural conditions as well as in various simulation devices, such an adjustment alone is not adequate. For this,

116

Weathering of Plastics

.•.

~ ~ ' ". ".

'" \.

--- .. --- --- ..f - - - - f - - - j - - - - j

30 i

'" 60 lrrad.,nce (300"'00 om) in W/m'

ST - surface temperature of different coloured samples

., I

""; tW< "" I

;'!lui

(Ia:l

:~-r~i

•

i I

--1

-j

L i

1,"-'-.--.......

~ ·1i~E(LM·8JWrr1

1-)UmirSnmn W-..-2·0Jdxr"'l='" - - - - .

. . • _• • ,

Figure 15. Surface temperature of PVC film having different colors.

1~ .J __

Temperature Measurement Techniques

117

fall systematically within the temperature field delineated by the black standard and white standard tenlperature. Note: A "calculated temperature field" with the conditions detailed in the figure key was "inserted behind" the measurement conditions indicated in ref. 3 by the identification of the equipnlent type, the black standard temperature, and the UV irradiation strength. A comparison of the surface telnperatures measured in the device with the test results fronl natural weathering (refer also to Figure 3) reveals an adequate degree of agreement between the measured values and those from exposure in the open at a 45° angle, facing south and without backing at a UV irradiation strength ofapprox. 60 W/m 2 (Figure 15, right side).

REPRODUCIBLE ADJUSTMENT OF TEMPERATURE CONDITIONS AT THE SAMPLE LEVEL IN WEATHERING EQUIPMENT Adequate agreement between the temperature fields with practical conditions on the one hand, and, on the other, basically identically designed weathering equipnlent - whose individual details nlay, however, vary - with xenon radiation sources, are determinants of both the correlation between, as well as the reproducibility of test results. In order to approach this goal, current equipment technology pragmatically utilizes the sample chamber temperature instead of the white standard temperature as the theoretical lower temperature limit, and uses this easily measured value as the regulated paranleter. Looking at the heat applied and removed at the surface of a given system - i.e., a black standard thermometer - the temperature can be seen to be a function of the irradiation strength, the temperature ofthe surrounding air (sample chamber temperature), and the velocity of the air stream at the sample surface. The temperature field illustrated in Figure 16 is displaced nearly in parallel with changes in the sample chanlber telnperature. If the air speed is altered, the field rotates around the (not indicated) ordinate intersection. It is therefore possible to adjust the black standard as well as the sample chamber temperature (and thus, by approximation, the white standard temperature as well) by varying the air velocity over a wide range in order to Inaintain a specified temperature field. Several American test standards (SAE J1860; SAE J1960; SAE J2212; SAE J2019) as well as the DIN 75 202 already specify a set temperature in the sample chamber along with specifications regarding the black standard/black panel temperature.

CONSTANT SURFACE TEMPERATURES AT INCREASED IRRADIATION STRENGTH (MEASURES TO INCREASE TIME-LAPSE) With regard to test standards, the application of increased UV irradiation strength in order to accelerate time-lapse effects is basically permitted under certain conditions. Such measures

Weathering of Plastics

118 .------------,r-----------.120 . 110

110 ;

100 ~--- .. ---_+_--+_"L-4__7.L--'

100 _.

\0'

r--=-'-b.-L:::;;;o-f--"='--=!=~...-1

~ ~

Cf!---+::=df

I-

i ·..··...... · I······......·..·

\\S'I~"" ".

i----+---r~~~+__=mn_i

~f..l__-_+'-O"-"'-!~d_

.'-------P-~>....;;t~,,:

-·~··iT'i·il

70 _ _-+_~l:_~*"-==-!!!!min!L

6J '.

50, min'

nin- nirirrunair spee1 rrax - treldrnm air spee1 30~:-----'-1---1-1- - - 4 -

o .

,,_..._ ..__ , ..

·f···..·······

3O'------'--~f----_+_-'--

._.

Figure f 6. Temperature fields in Xenotest.

-! L

00 -'

have, for several years, been actively discussed and examined within the context of illumination and weathering tests for materials intended for automobiles. Regardless of their admissibility and the limits of such parameter intensification with respect to the aging process, the relationship with a simultaneous alteration

in the surface temperature level must be carefully considered. To a limited extent, the UV ilTadiation strength can be increased without altering the temperature values for the BST and WST by adjusting the velocity of the air flow. An additional increase is linked to an alteration of the limit temperatures and thus the surface temperatures of samples. In order to maintain an approximately unaltered temperature profile under these circ*mstances, the IR portion of the xenon radiation must be reduced, while the spectrum itself remains unchanged. Currently, this is achieved by means of an appropriate filter system or, in the case of the Alpha xenon test device, by means ofa modified xenon lamp whose infrared radiation portion has been reduced. The latter technique in particular has opened the way for testing with very high UV ilTadiation levels at unchanged BST and WST levels. 13-15 However, the presence ofan influence on the final surface temperatures as a result of the altered spectral distribution must be borne in mind on a case-by-case basis.

SUMMARY The effect of solar radiation on materials results in primary photochemical processes and secondary decomposition processes. The later are significantly influenced by other climatic parameters. Of these, the material-specific final surface temperature that is directly determined by the solar radiation has special significance. The measurement techniques to characterize the surface temperatures of samples in weathering technology have been discussed in detail.

Temperature Measurement Techniques

119

The risk of an inadequate indication of surface temperatures of exposed materials inherent in the use of "random" black panel thermometers has been expounded. The principle and design of the improved black standard them10meter measuring element have been indicated, and the systemic differences of the temperature values given by the BPT and BST have been discussed. Examples have been used to demonstrate that, with the utilization of black and white standard in current equipment, it is possible: • To create or regulate identical temperature conditions in different types of xenon weathering equipment; • To adequately adjust the final surface temperatures achieved under the influence of solar radiation in the open-air in equipment employing filtered xenon radiation. Thus, a further step towards the goal of in1proved reproducibility and correlation of weathering tests has been taken.

REFERENCES 2 3 4 5 6 7 8 9 10 11 12

14 15

VOl 3789; Blatt 2. Umweltmeteorologie; Wechselwirkungen zwischen Atmosphaere und Oberflaechen; Berechnung der kurz- und lang\velligen Strahlung; October, 1994. CIE No. 85. Technical Report; Solar Spectrallrradiance; Table 4; 1989. Fischer, R.M. and Ketola, W.O. Surface Temperatures of Materials in Exterior Exposures and Artificial Accelerated Tests; Accelerated and Outdoor Durability Testing of Organic Materials, ASTM, STP 1202, 1994. Schulz, U. Der Einfluss von Temperatur und Feuchte auf die photochemische Alterung polyn1erer Werkstoffe, Seminar Nr. 510235004 - Natueliches und kuenstliches Bewittern polymerer Werkstoffe; TA Wuppertal, 1994. Wypych, 1. '''eathering Handbook, Chem Tee Publishing, Toronto 1990, p. 41. Boxhammer, 1., Kockott, D., and Trubiroha, P. Black Standard Thermometer - Temperature Measurement of Polymer Surfaces During Weathering Tests; Materialpruejung, 35 (1993) 5; p. 143-147. Boxhammer, J. Temperaturmessung in der Ebene exponierter Proben bei der zeitraffenden Bestrahlung/Bewitterung in Geraeten; Seminar Nr. 510235004 - Natueliches und kuenstliches Bewittern polytnerer Werkstoffe; TA Wuppertal, 1994. DIN 53387 Ausgabe 1989; Pruefung von Kunststoffen und Elastomeren; Kuenstliches Bewittern und Bestrahlen in Geraeten - Beanspruchung durch gefilterte Xenonstrahlung. DIN 75 202 Ausgabe 1988; Bestimmung der Lichtechtheit von Werkstoffen der KraftfahrzeuginnenauBS Tattung - Xenonstrahlung; Anm: Revidierte Fassung liegt vor, ist jedoch noch nicht publiziert. ISO 4892 Ausgabe 1994; Plastics - Methods ofExposure to Laboratory Light Sources, Part 1: General Guidance and Part 2: Xenon Arc Sources. ISO/TC611SC6/WG2 - Light Exposure; Task group: Comparison ofBPT and BST; results as yet not published. Boxhamtner, 1. Current Status of Light and Weather Fastness Standards - New Equipment Technologies, Operating Procedures and Application of Standard Reference Materials; Material Life Society; 2nd International Symposium on Weatherability, Tokyo, September, 1994. Boxhammer, 1. Einfuehrung eines neuen Temperatunnesssystems zur Verbesserung der Reproduzierbarkeit von Be\vitterungsversuchen; Sen1inar Nr. 102019 - Bestrahlen und Bewittern von polyn1eren Werkstoffen; TA Wuppertal, 1989. Crewdson, L. F. E. Correlation of Outdoor and Laboratory Accelerated Weathering Tests at Currently and Higher Irradiance Levels - Part III; Material Life Society, pg. L /13/. Boxhamlner, 1. A. COlnparison of New and Established Accelerated Weathering Devices in Aging Studies of Polymeric Materials at Elevated Irradiance and Temperature; Material Life Society; 3rd International Symposium on Weatherability, Tokyo, May, 1997.

Infrared Welding of Thermoplastics: Characterization of Transmission Behavior of Eleven Thermoplastics

Hong Jun Yeh and Robert A Grimm Edison Welding Institute

INTRODUCTION Through-transmission infrared welding (TTIR) of thermoplastics has been detailed in a previous paper. 1 In general terms, it involves heating ofthe weld zone by transmission ofinfrared energy through a transmitting polymer and onto an absorbing polymer that is in contact with it. The aforementioned paper addressed welding of clear acrylic (PMMA) as the transmitting material to black PC as the absorbing material. The focus of the present paper is to explore a range of COlnmon polymers to determine their suitability for TTIR. Quartz-halogen lamps, with filament temperatures in the range of3000oe, emit radiation over a range of wavelengths but the maximum output is at a wavelength predicted by Wein's Law. 2

AmaxT ==O.2898x10 4 Jlm oK For example, a quartz-halogen lamp with a filament temperature of 30000 e (3273°K) has a maximum output at 0.89 11m. A rod heater operating at 800 0 e (1073°K) has a maximum output at 2.7 Jlm. A heated platen on a hot plate at 400 0 e (673°K) has a maximum output at 4.3 Jlm. While the maxitnum output is at the specified wavelengths, the actual output is a distribution. With the quartz-halogen lamp, for example, slnall amounts are emitted at wavelengths as long as 5 Jlm and as short as 0.3 Jlln (ultraviolet rays). A typical distribution curve is shown in Figure 1. Radiation can be either reflected from the surface of the polymer substrate, absorbed by the bulk of the polymer, or transnlitted through the substrate.

122

Weathering of Plastics

3300'K Filament Temp

Natural, unfilled polymers absorb infrared radiation at specific frequencies that are characteristic of the molecular structure of 80 the polymer. The most common type of infrared spectrum spans the frequency range from 1.5 to 15 j..lm and, within this range, there are specific absorption bands that are due to stretching, rocking, scissoring, and .... similar types of molecular motion. Since the energy levels at which these motions occur are quantized, they occur in relatively naro row bands. Most polymers will absorb o I 2 wavelengths from 3.2 to 3.6j..lill because they UV Vis Near IR Middle lR contain carbon-hydrogen bonds. If polymers contain alcohol, carboxylic acid, or amide Walckngth [microns) groups, absorption bands (often broad) are Figure I. Approximate spectral distribution for a quartz-halogen seen around 2 to 3 j..lm. Between the 3.6 j..lm lamp. bands and around 6 to 7 j..lm, most polymers are relatively transparent. The near-infrared part of the spectrum ranges from the red part of the visible spectrum (0. 72j..lm) to the 1.5j..lm wavelength described above. This part of the spectrum does not have as much utility in characterizing polymers because it contains relatively weak absorptions or overtones of primary absorptions that appear in the midrange infrared spectrum. lOll

""'"

EXPERIMENTAL DESCRIPTION OF EQUIPMENT AND PROCESS Absorption behavior was inferred by measuring the amount of energy that was transmitted through the polymer films. The configuration of the lamp, film, and measuring devices is shown in Figure 2. Since the output from the lamp was constant, the greater the transmission, the less the absorption. This type of arrangement provides qualitative infonnation that can be used for comparative purposes. Prolonged exposure led to melting and decomposition of the films, so care was used to expose the samples for lengths of time where they were not damaged. The amount of transmitted energy was measured with a fixed thermocouple or with a radiant energy power meter. Output was recorded with a computer-controlled data acquisition system. The sensor was positioned on a block of white polymer (polyoxymethylene) to avoid

Characterization of Transmission Behavior

123

any charring in the region of the sensor since this can have a significant effect on temperature readings. The spot heating lamp was a custom-built, MR16, Infrared quartz-halogen type (General Electric EXS with a focal Lamp length of4 cm and a smooth, alulninized reflector) operating at filament temperature around 3000°C. At the focal plane, the spot heater delivered a flux density in excess of 140 W/cln2 • At these temperatures, the nlaximum output occurred at a wavelength around 0.89 JJ1l1. Infrared radiation is not visible to the human eye, but these lanlps emit visible light of considerable intenPolymer sity along with small amounts of ultraviolet light. For ---------Film this reason, protective, dark green glasses were worn during all tests and EWI recommends this as a standard safety practice. "

MATERIALS Figure 2. Experimental set-up.

Pellets of natural, unfilled thermoplastics were pressed into films by melting and pressing them between two sheets of Kapton@ film in a platen press. After pressing, the films were cooled and the Kapton'ID peeled. Film thicknesses were controlled by placing shinls between the platens and were measured with a micrometer. The polymers included PS, PMMA, PC, HDPE, PP, PA-6, ABS, PPS, EVAI, PVC, HIPS, PA-6 fOlmed a clear film under these conditions. When thermocouples were used, readings were taken at a 2-second heating time. A few runs were made to test reproducibility and it was found to be reasonably good (PC - 270 and 275°C, PP - 235 and 240°C).

RESULTS AND DISCUSSION TEMPERATURE VS. THICKNESS AND POLYMER TYPE

All of the polymer films that were tested show substantial amounts of transmission, as would be predicted by the generalization that the absorption bands in the region around 1 Jlm are relatively weak. In separate tests, it was determined that the transmission through air was only about 10% better than through the transparent polymers.

124

Weathering of Plastics

.w'l------

...

Ilf)

!'-'~""

¥ __•

••••••• ~

••

~-!

, -.- PPS

-H-AUS

-+-I'S

f'llH'

-'6-lH)U

If)

'" I

O)05mtll

O.6lllm

091

m{ll

7.2 t1~1l

Figure 3. Graph of transmitted energy vs. film thickness for several translucent polymers.

Figure 4. Graph of transmitted energy vs. film thickness for several transparent polymers (note the shallower slope with these materials).

Figure 3 shows the transmission behavior for several translucent polymers. These are normally opaque in the visible range because they are crystalline or contain insoluble phases (rubber in ABS or impact modifiers in HIPS). Polymers that contain high levels ofC-H bonds or which have extra bands in the region shorter than 5/lill, such as the N-H bonds in polyamides or the CN bonds in ABS seem to absorb a little more radiation than the others. Various thicknesses of the polymers were examined for transmission of infrared radiation. With polyethylene, for example, a single layer allowed the thermocouple to reach 275°C while the second layer dropped this temperature by 40 to 235°C. Addition of the third layer resulted in another 40°C drop to 195°C while the fourth layer only caused a 20°C decrease to l75°C. This behavior appeared to be general. The translucent or opaque polymers show curves that span a range of initial temperatures (thinnest samples). In all cases (except for PPS with only two points), the temperature curves show a slight upward curvature.

Characterization of Transmission Behavior

125

This observation can be rationalized on the basis that the alTIOunt of energy that is transmitted is a function of two effects: one is the background absorption (scattering) since none of these polymers is completely transparent, the second is the small absorption bands that are present in the near infrared part of the spectrum. The background absorption will be additive and unavoidable for each layer. However, the energy absorption due to the small bands in the near infrared part of the spectrum will occur primarily in the top layer which acts as a filter. Once these wavelengths are removed, they are not present to heat the underlying layers, thus leading to an upward curvature in the thickness graph. The materials showing the lowest transmission are ABS and HIPS, both contain impact modifiers (rubber particles). The conclusion is that the particles increasingly scatter the infrared radiation compared to polymers that simply contain crystallites. Next to these two is PVC, which is normally fOlIDulated with stabilizers to prevent decomposition. These are often inorganic Inaterials that are truly opaque to all wavelengths of light. Polyethylene and polypropylene are opaque to visible light but are relatively transparent to the infrared. They are clean polymers with essentially no opaque additives. Still, they are not quite as transparent as the clear polymers, probably due to the high levels ofC-H bonds in these molecules that absorb around 3.4 J.lm. Figure 4 shows the curves for the transparent polymers. One observation is that transmission is higher, yielding temperatures of 280 to 300°C for the thin levels. By contrast, the translucent polymers transmit less, yielding temperatures of 200 to 275°C for comparable thicknesses. Another observation is that there is a smaller decrease in transmission as the polymer thickness increases. While the translucent polymers show a decline in transmitted temperatures of 80 to 100°C as thicknesses increase from 0.3 to 1.2 nlm, this decline is only about 50 or 60°C for the transparent polymers. The exception to all these observations is EVAl. Even though this is an optically clear polymer, transmission is lower for the thin levels and the decrease with thickness is more like that of the translucent polymers. A rationalization for this behavior is that EVAI is rich in both hydroxyl groups (OH) and carbon hydrogen bonds (C-H). Both of these structures absorb strongly around 2.5 to 3.5 ~l1n.

CONCLUSIONS The data shows that all of the parent, unpigmented polymers examined will transmit infrared radiation at the wavelengths produced by a quartz-halogen lanlp. Opaque polynlers, such as PP, HDPE, PVC, PPS, ABS, and HIPS transmit slightly less efficiently than clear ones such as polycarbonate, but all these unfilled, natural polymers transmitted the majority of the incident radiation.

126

Weathering of Plastics

REFERENCES 2

Grimm, R. A., Through-Translnission Infrared Welding of Polymers, Conference Proceedings ofthe SPE ANTEC, 1996, Indianapolis, IN., p. 1238. Ference, M., Lemon, H., and Stephenson, R., Analytical and Experimental Physics, Second Edition, The University of Chicago Press, Chicago, IL, 1956, p. 550, and similar physics or optics textbooks.

Infrared Welding of Thermoplastics. Colored Pigments and Carbon Black Levels on Transmission of Infrared Radiation

Robert A Grimm and Hong Yeh Edison Welding Institute

INTRODUCTION Infrared welding has been characterized as unpredictable since different polymers or formulations have been observed to heat at widely different rates under similar conditions. A previous reference reported a significant difference in absorption between thermoplastics that contained no carbon black and a similar material with carbon black levels around 0.5 percent,l but no intermediate levels were examined to determine the minimunl levels at which absorption became essentially total. Polymers that have pigments with other colors of the spectrum and/or can scatter light might also be expected to show differences in absorption of infrared energy. This is because the flux density of the radiation from a nonnal quartz-halogen source is greater on the red side of the spectrum than it is at the blue end of the spectrum. Quartz-halogen lalnps, with filament temperatures in the range of3000°C, heat predominantly through radiation. The output from filament or thermally-heated sources can span a range ofwavelengths but the n1aximum output is at a wavelength predicted by Wien's Law. 2

While most of the radiation from a quartz-halogen source is emitted at 0.89 Jlm, small amounts are emitted at wavelengths as short as 0.3 Jlm (ultraviolet rays). There is a substantial visible component to this light and there is relatively more red than blue in it. A distribution curve is shown in Figure 1, but it should only be considered approximate at this tilne.

128

Weathering of Plastics

]]fjrK Ebmenl Teo)) 1fJ()

o ! UV Vis Nw IR Middle IR Waveleog:n (micronsl

Pigments, fillers, coatings, and other components of polymer formulations affect what happens to incident infrared radiation. They will strongly affect the ability of polymers to absorb the IR radiation, and can lead to various amounts of reflection and absorption (and transmission). This study aimed to study the absorption characteristics of IR energy by different polymers, by variously-colored ABS materials, and by polymers with different levels of carbon black. This information should provide practical guidelines and process understanding when infrared welding is being used.

EXPERIMENTAL DESCRIPTION OF EQUIPMENT AND PROCESS

Films of colored polymers were obtained by disassembling floppy diskettes (ABS) of different colors and using segments of the O.25-mm-thick walls. Red, orange, yellow, green, and blue pieces were examined with thicknesses of 0.25 and 0.5 mm (two layers). The polyethylene films with various levels of carbon black were prepared by mixing various ratios of black polyethylene (0.2 percent C) with natural polyethylene (w/w) and pressing them between Kapton1\) films in a platen press. The film was cut, restacked, and re-pressed about ten times. After these multiple pressings, uniform films were obtained in thicknesses of approximately 0.25 and 0.5 mm. Films were prepared with carbon black levels of 0.033, 0.05, 0.067, 0.1,0.133, and 0.15 percent carbon. Absorption behavior was inferred by measuring the amount ofenergy that was transmitted through the polymer films. The configuration of the lamp, film, and measuring devices is shown in Figure 2. Since the output from the lamp was constant, the greater the transmission, the less the absorption. This type of arrangement provides qualitative information that can be used for comparative purposes. Prolonged exposure led to melting and decomposition of the films, so care was used to expose the samples for lengths of time where they were not damaged. Figure I. Nonnalized spectral output from a quartz-halogen lamp (shaded area is the visible and UV range(O.3 to 0.72 J.l m)).

129

Transmission of Infrared Radiation

Lamp

! •.•. ,

80 70

• •

._.•.....

0.25 mm 0.5 mm -

.~._

.. _--.

--_.

60 ----U .....

Film

I-<

....;::l ~ l-

0..

Figure 2. Experimental set-up.

10 0

The amount of transmitted energy was measured with a fixed thermocouple. Output was recorded with a computer-controlled data acquisition system operating with a sampling rate of 10 Hz. The thermocouple was posi _

Carbon (%) Figure 3. Effect of carbon black level on transmission through polyethylene.

tioned on a block of white acetal polymer (polyoxymethylene) to avoid any charring in the region of the sensor since this can have a significant effect on temperature readings. The spot heating lamp was a custom-built, MR16, quattz-halogen type (General Electric EXS with a focal length of 4 cm and a smooth, aluminized reflector) operating at filament temperature around 3000°C. At the focal plane, the spot heater delivered a flux density in excess of 140 W/cm 2• At these temperatures, the maximum output occulTed at a wavelength around 0.89 11m. Infrared radiation is not visible to the human eye, but these lamps emit visible light of considerable intensity along with small amounts of ultraviolet light. For this reason, protective, dark green glasses were worn during all tests and EWI recommends this as a standard safety practice.

130

Weathering of Plastics

RESULTS AND DISCUSSION 100

EFFECTS OF CARBON BLACK LEVEL

Carbon black pigments found in black polymers represent an almost perfect absorbing material. The work reported here aimed at quantifying how light transmission depended on the level ofcarbon black that is present in a polymer. 40 Figure 3 shows the findings. The thin film of 0.25 mm showed a definite S-shaped variation, going from transparency to nearly complete absorption as the carbon black level went from 0.03 to 0.07 percent. Ifpolymers are formulated with carbon black levels os o in this range, their welding beTime (sec) havior can be expected to be Figure 4. Graph showing transmission vs. color of ASS sheets. highly sensitive to tiny variations in formulation. The thicker film, at 0.5 mm, shows much less ofa transition in the range of 0.03 to 0.07 percent carbon. At this thickness, the effective amount of carbon black is increased, essentially doubled, and transparency will require even lower levels. These conclusions have been confirmed in studies conducted since this work was completed. 80

EFFECTS OF COLORED PIGMENTS ON ABSORPTION

ABS samples were cut from the walls of colored floppy disks and used as single and double layers for evaluation of the effects of color on absorption. It was assumed that the polymer formulations were the same, except for the color, thus isolating the effects of part color. Heating due to the near infrared wavelengths should have remained constant. In summary, absorption was strongest for the red samples, weakest for the blue, and decreased as the colors changed from yellow to orange to green (Figure 4). Examination of the

Transmission of Infrared Radiation

131

output spectrum from a quartz-halogen lamp shows that it has a higher flux density at the red end of the spectrum rather than the blue since output goes to zero just beyond the blue or violet end of the spectrum. Materials have a certain color because they absorb all colors from the spectrum and re-emit the observed color. Thus, the red sample should be re-emitting the red color that constitutes the largest part of the visible output. However, the temperatures sensed with the red samples are lower than for other colors. Thus less energy is transmitted through the red ABS than through any other color. The blue polynler is re-emitting the blue color which constitutes only a small part of the visible output. However, it results in a higher temperature reading by the thennocouple, indicating a higher level of energy transmission.

CONCLUSIONS This work shows that absorption is very sensitive to the level of carbon black in the polymer fonnulation and provides some data to quantify this effect. This sensitivity occurs at very low levels of carbon black. Thus, when a polYlner is selected for infrared welding, it will be important to know the concentration of carbon black in the fOffilulation. If it falls below 0.07 percent, there will be increasing depth of heating and less surface heating. Levels in excess of 0.03 percent carbon will heat primarily by surface absorption of the infrared radiation. In this latter case, the creation of a significant depth of melting will depend on the relatively slow process ofconduction. However, if changeover times are short, this latter method will approach high temperature hot plate welding where surface decomposition is tolerated so long as the decomposed material is squeezed out as flash. One significant risk is that the joints may not be as strong as when a deeper melt zone is created. Because infrared welding has the ability to penetrate polymers and heat them, it offers the potential for stronger joints because a deeper lnelt zone is created by absorption, at once, rather than by conduction through the polymer. Plastic parts with a thin black layer on one side can be continuously welded in place. For example, if the black layer is one part of a bi-Iayer, coextruded sheet, it could be unrolled and welded in place. The heat needed for welding would be generated precisely where it is needed, minimizing damage to the part and allowing the joining of thin polymer fihns. Thin films, particularly when coated on a metal, that are hard to join by other methods should be readily and rapidly joined by infrared welding. Polymers of different colors can be expected to weld differently by infrared welding. Not only are the issues of pignlent-polymer interactions present such as the differences in weldability caused by white (titanium dioxide), black (carbon black) or other pigments, but heating times and depth of heating are likely to be affected by part color. This kind of phe-

132

Weathering of Plastics

nomenon can already occur in conventional hot plate welding when, in some cases, red and black parts weld differently. It can be expected to beconle even more ofan issue with infrared welding. These effects can be easily handled, but workers must be aware oftheir presence and how to control them.

REFERENCES 2

Grimm, R. A., Through-Transmission Infrared Welding ofPolymers, Conference Proceedings ofthe SPE ANTEC, 1996, Indianapolis, IN., p. 1238. Chen, Y. S. and Benatar, A., Infrared Welding ofPolybutylene Terephthalate, Conference Proceedings ofthe SPE ANTEC, 1995, Boston, MA., p. 1248. Ference, M., Lemon, H., and Stephenson, R., Analytical and Experimental Physics, Second Edition, The University of Chicago Press, Chicago, IL, 1956, p. 550, and similar physics or optics textbooks.

Predicting Maximum Field Service Temperatures From Solar Reflectance Measurements of Vinyl

Henry K. Hardcastle III Da)Jton Technologies, USA

INTRODUCTION A number of vinyl building product manufacturers are familiar with The Standard Test Method for Predicting Heat Buildup in PVC Building Products according to ASTM D 4803 which utilizes an insulated box to house a specimen irradiated by an IR heat lamp. Many vinyl producers may not be familiar with the basis of this test or the direct measurenlents that can be tnade to predict the propensity for heat buildup.! Recent failures of rigid vinyl materials due to heat buildup and heat distortions have been observed even though ASTM D-4803 analysis indicate acceptable performance. These materials have also displayed satisfactory heat buildup performance in historical markets. Sales and subsequent failures of these products in newer Western US markets itnply an environmental constraint not found in traditional eastern geography's and a possible limitation to the D-4803 method. Failures that initiated this study have been focused around areas with higher solar irradiance in the Southwestern US.

STATEMENT OF THEORY AND DEFINITIONS THE SOLAR SPECTRUM The solar spectrum is a depiction of the energy from the sun that irradiates a nlaterial. Due to filtering effects of the atmosphere more than 98% of the sun's energy that strike the earth's surface are between 300 and 2500 nm. The radiant energy at any particular wave band within this spectrunl is highly dependent on the amount and quality of atmosphere the energy travels through before striking the material. • There are several different agreed upon solar spectrums. • One of the major differences is the amount of attnosphere the energy must travel through.

Weathering of Plastics

134

Another difference is the amount of direct vs. diffuse light irradiating the 2001) surface. •E o Three major solar spectmms defined ~ 1500 8 are Air Mass 1.5 Direct, Air Mass 1.5 Ii 1000 Global and Air Mass 0 as shown in ~ 500 Figure 1. o There may be other sources of irradiance besides the sun contributing to heat build including; shingles Wavelength (nm) that are reflecting or re-radiating at Figure I. Three different ASTM standard solar spectrums. long wavelengths, low E glass, bar-b-que grills, pool decks and other good absorbers, emitters or reflectors of solar energy. Often these features may concentrate solar energy or re-radiate absorbed solar energy at longer wavelengths and contribute additional energy for heat buildup. 2500 ;--.-...--...- - - - - - - - - - - - - - - - - - ,

E 490 Air Mass 0

VINYL OPTICAL PROPERTIES Optical properties can be characterized using the relationship: 1=p+'t+u

[1 ]

The relationship simply states that the total irradiance striking a material will either be reflected off the material, transmitted through the material or absorbed by the material. It is the absorbed solar energy that is available for heat buildup.2 The relationship becomes even more simple if the material is opaque (r = 0). It is important to consider a materials optical properties through out the entire solar spectmm (approximately 300 to 2500 nm) rather than just the visible spectmm or just the IR spectmm since about half of the solar energy is composed of wavelengths less than 780 nm and half the solar energy lies above 780 nm. Some materials that have low absorptance in the visible portion of the solar spectmm may have high absorptance in the IR region. Pigment manufacturer's take advantage of this fact and produce many products often referred to as "IR reflective pigments" that appear dark in visible light but are highly reflective in the IR and therefor remain cooler than similar colors made with traditional pigments.

135

Maximum Field Service Temperatures

DESCRIPTION OF EQUIPMENT AND PROCESSES MEASUREMENT OF OPTICAL PROPERTIES Measurement of reflectance and transmittance optical properties is easily accomplished using modem commercially available spectrophotometers. • It is important that the spectrophotometer has the ability to scan the majority of the solar spectrum from approximately 300 to 2500 om. • The geometry of the measurement, incident and reflected angle of spectrophotometer beams, reference beams and use of integrating spheres are important considerations of these measurements especially when comparing optical properties measured using different configurations or instruments. • Measurement geometry and front end optical designs are well documented in ASTM E903 for these measurements. • The initial results of these optical properties measurements is typically a spectral reflectance or transmittance curve showing the %p or %r at each wavelength as a graph as shown in Figure 2. 100 90

:l u

~ ~

; '"

eo . 70 60

so ~o

30 20

to 0 0

;;

• 0

~ ~

000

~ ~

000

§ ~ § $

000

N ~ ~ ~

Wive Llngth (nm)

Figure 2. Spectral reflectance curves for 3 colors of rigid vinyl.

INTEGRATION OF OPTICAL PROPERTIES TO THE SOLAR SPECTRUM Once a measurement of the percent reflectance and percent transmittance of the material at each wavelength is obtained through out the solar spectral region (300 - 2500 nm) the optical properties of the material may be related to the sun's irradiance by integration. Integration is a mathematical weighting process that takes into account both the sun's irradiance and the material's reflectance at each wavelength from 300 to 2500 om. Integration weights regions of the material's optical properties spectrum according to the energy output from the sun in those regions. • Once the sun's irradiance and material's optical properties are integrated at each wavelength, the total of reflected solar energy may be summed resulting in a single I

136

Weathering of Plastics

number denoted as "total percent solar reflectance" for the air mass used. Percent solar absorptance is then calculated:

a == 1- (p + 1: )

[2]

• It is the value of percent solar reflectance and the calculated percent solar absorptance that is powerful in predicting a materials propensity for heat build. • For opaque tnaterials such as a rigid vinyl, colors with high solar reflectance will remain cooler than colors with low solar reflectance under the same environmental conditions. • For materials with the same emittance characteristics, materials with higher solar absorptance will have a greater propensity for heat build. Materials with lower solar absorptance should remain cooler for similar materials under the same solar and ambient conditions.

APPLICATION OF EQUIPMENT AND PROCESSES There appear to be 4 main steps to using the solar spectrum, optical property measurements and solar integration; 1 Define the temperature failure criteria for the material. 2 Obtain empirical heat build up data for a number of material colors in worst case environments. 3 Measure the optical properties of the material colors and plot correlation regression between solar absorptance and worst case empirical heat build data noting where the regression line crosses the failure criteria. 4 Consider the risks involved with selling products which measure above the critical solar absorptance characterized in the previous step. An example will demonstrate use of these four steps. EXAMPLE OF METHODOLOGY

A producer offers a variety ofdifferent colors in the same PVC base. Colors are formulated by altering the pigments and Ti0 2 content. In this example, the producer has no prior knowledge of field performance but wants to determine the maximum solar absorptance he can design and still have acceptable heat buildup performance. 1) Define the temperature failure criteria for the material. The producer determines experimentally the maximum service temperature his material can achieve and still provide acceptable performance. The producer determines the heat deflection temperature (ASTM D 648), Vicat Softening Temperature (ASTM D 1525), Coefficient of Thermal Expansion

137

Maximum Field Service Temperatures

(ASTM D 696) or other appropriate quantitative measures of material's performance under heat. The producer then adds a suitable safety factor to the temperature determined to cause failure. 2) Obtain empirical heat build up data for a number of material colors in worst case environments. The producer obtains a nurrlber of samples of different colors of his material and exposes them to the worst case environment in his intended market. This environment should have the highest solar irradiance and wamlest temperatures the product may be subjected to while in service. The samples should be oriented for exposure resulting in the maxinlum heat build; oriented nomlal to sun, protected from breezes and insulated from convective and conductive cooling as much as appropriate for the product. Consideration should also be given to reflective surfaces and other heat sources the product may encounter in the field. The producer then carefully measures the temperatures the selected samples reach under these worst case conditions using thermocouples, pyrometers or other suitable temperature measuring and data logging instrunlentation. The temperature nleasurements are made simultaneously for all specimens to block differences in enVir01l111entai variables as shown in Figure 3. 90

6 rown

BO 0

:::l

fij

R E III

60 50

Ambient

40 30

a; C'.J

Oi I,{} 0

~ 0; ~ y-

8m ¥':!

l¥

If)

1.0 0

&; l.() N

V> 0

(0';

(\1

f.1)

ig m

0) N

l.O 0

Local Tim e

Figure 3. Heat buildup near smntner solstice in Phoenix, AZ for 3 colors of rigid vinyl.

3) Measure the optical properties of the material colors and plot cOlTelation regression between solar absorptance and worst case empirical heat build data noting where the regression line crosses the failure criteria. The producer then measures the solar optical prope11ies of the samples used to obtain the worst case heat build temperatures and calculates solar absorptance. An x-y scatter plot is then constructed with maximum temperature on the ordinate and solar absorptance on the abscissa. The regression line is fitted to the data. The temperature failure criteria from step 2 is marked on the ordinate scale and a line is extended to intersect with the regression line as shown in Figure 4. The point of intersection with the re-

Weathering of Plastics

138

50 . r

Critical Ab$orptance

--~-~-----~---r----~~----~--~~---+-----+------4 10 3D 40 20 100 70 80 90 50 50

+""'.

()

Solar Abs orptanc:e

Figure 4. Maximum temperature vs. solar absorptance scatter plot denoting critical absorptance for a rigid vinyl system.

gression line is then extended down to the abscissa and solar absorptance indicated becomes the design criteria for new products. 4) Consider the risks involved with selling products which measure above the critical solar absorptance characterized in the previous step. Products made with higher solar absorptances have a higher risk of exceeding the defined temperature failure criteria determined in step 1. Again, the producer may choose to utilize IR reflecting pigments to produce dark colors or decide the cost and risk outweigh the revenues from color offerings with higher solar absorptance.

PRESENTATION OF DATA AND RESULTS ACTUAL CASE STUDY DATA The company had been selling traditional colors of window lineals since the mid 1980's with acceptable heat build performance. Colors sold included white, beige and brown in standard rigid vinyl formulations. The company utilized ASTM D 4803 to evaluate heat buildup before a new color's introduction. In the past several years, however, several things happened that increased risk of heat related failures: 1 Markets moved westward from northeastern and southeastern and central US areas to southwestern US markets such as Denver, Las Vegas, Phoenix, Southern California, etc. 2 Customers began demanding new custon1 colors including very dark colors.

139

Maximum Field Service Temperatures

The company began experimenting with new formulations including different types and amounts of lubricants, stabilizers, impact modifiers, and colorant vehicles. 4 Heat related problelTIs became a focus of discussions alTIOng building product producers and suppliers. The challenge for the R&D effort was to develop a methodology to predict propensity for heat buildup for new experimental formulations in addition to the D4803 method. The new method needed to be empirically based and applicable to the data base ofperformance already available (e.g., customer complaints and historical product offerings). Finally, this method needed to provide decision makers with a clear indication of new products performance before release to the markets. 1) Define the temperature failure criteria for the material. The experimental formulas were blended and extruded. The extruded products were measured for heat deflection temperature using ASTM D 648 as a guideline. Multiple measurements at various heating rates were conducted. An appropriate engineering safety factor was applied to the data. A critical temperature failure criteria was defined as 70°C for these particular experimental formulas. 70°C was considered the maximum sustained temperature the extrusions could withstand and still provide acceptable engineering performance. 2) Obtain empirical heat build up data for a number ofmaterial colors in worst case environments. A collection of 11 specimens representing the range of current product offerings and R&D efforts was selected. The materials were mounted in a single standard frame, side by side. The specimens were similar in thickness and dimension. The frame and specimens were mounted over standard building insulation to prevent back side cooling and surrounded by wind baffles to reduce cooling due to breezes. ThelIDocouples attached specimens to a simple data logger. The specimens were exposed directly to sun at Phoenix, AZ near summer solstice 1997 at near nonnal angles. Measurements were taken continuously for several days. The maximum temperature achieved by all specimens at the same time was recorded. An exalnple of this data is shown in Figure 3. These values were then described as the best estimate of heat buildup for the colors in a worst case environment. 3) Measure the optical properties of the material colors and plot con elation regression between solar absorptance and worst case empirical heat build data noting where the regression line crosses the failure criteria. Each of the materials was then measured using ASTM E 903 and integrated using ASTM E 892. Each material was opaque. The percent solar absorptance was calculated for each material. The solar absorptance vs. maximum heat build were plotted in x-y scatter plot format and fitted with a regression line. The tenlperature failure criteria was noted on the temperature scale and extended to the regression line. The point of intersection denoted the maximum % solar absorptance that could be achieved by the system and still provide acceptable heat buildup performance as shown in Figure 4. For this formulation, the maxinlum solar absorptance should not exceed 40% a critical value. 4

140

Weathering of Plastics

4) Consider the risks involved with selling products which measure below the critical solar reflectance characterized in the previous step. The critical solar absorptance value of40% became a clear design criteria for current and new color product offerings in this system.

INTERPRETATION OF DATA The empirically derived maximum temperature vs. solar absorptance regression shown in Figure 4 becomes an important tool for new product designers using this vinyl system. Different colors produced in this fonnulation can be identified on the regression by simply measuring their solar reflectance and calculating their solar absorptance value. Once a custom color is nlatched, a sample is immediately submitted for solar reflectance measurements. If a pigmentation system used to achieve a custom color results in solar absorptance values above the critical value, decision makers know the probability of heat related complaints will increase in severe environments.

SUMMARY AND CONCLUSIONS Use ofempirically derived heat buildup data and optical properties measurements can significantly inlprove a producer's ability to predict maximum field service temperatures of vinyl materials. Use of empirical field methods described here in addition to laboratory tests can identify robust design criteria, enhance a product's service performance and ultimately contribute to customer satisfaction.

ACKNOWLEDGMENT The Author would like to acknowledge Dayton Technologies for permission to publish this work.

REFERENCES 1 2

B.B. Rabinovitch, et al. 1. Vinyl Tech., 5. No.3 (1983). Duffie, lA. and W.A. Beckman, Solar Engineering of Thermal Processes. John Wiley and Sons, 1980, p. 144-154.

Residual Stress Distribution Modification Caused by Weathering

Li Tong and J R White Materials Division, University ofNewcastle, Newcastle upon Tyne NEI 7RU, UK

INTRODUCTION Residual stresses fonn in thertTIoplastic moldings as the result of the temperature gradients present during solidification. 1-3 During subsequent service the stress magnitudes may beconle reduced as the result ofrelaxation4 ,s while in some circ*mstances more complex changes may occur. Paterson and White have examined the effects ofwater absorption into Nylon 66, causing swelling, secondary crystallization and a change in modulus. 6 ,7 In other studies Qayyum and White have observed significant changes in residual stress distributions in injection molded bars exposed outdoors in a hot sunny climate. 8,9 These changes were tentatively attributed to the presence of temperature gradients at certain times of the day, and similar effects were reproduced in the laboratory. 10,1 1 Chain scission caused by photo-oxidation may lead to secondary crystallization with crystallizing polymers; this is often called chemi-crystallization. I2 ,I3 Secondary crystallization is strongest near the surface of the molding and varies markedly through the depth. This will cause changes in the residual stress distribution independent of any effects of relaxation caused by temperature gradients. The current paper shows how the changes in crystallinity can be used to predict the changes in residual stress distribution and compares the results with measurements made using the layer removal procedure. 1-3

PREDICTION OF CHANGES IN RESIDUAL STRESSES CAUSED BY CHEMI-CRYSTALLIZATION The analysis will be confined to parallel-sided moldings. It is first necessary to consider whether the changes will occur symmetrically. If the molding is exposed outdoors and receives sunlight equally on both surfaces it can be assumed that the changes will be symmetri-

142

Weathering of Plastics

cal about the mid-plane. This will also be true if the molding is illuminated equally on both sides in the laboratory. The symmetrical case is considered below. If, on the other hand, the molding always presents the same surface to the sun or to the source of illumination in the laboratory then a different pattern of chemi-crystallization will develop near the exposed and the unexposed surfaces respectively. The limiting case will be that in which there is no chemi-crystallization near to the unexposed surface and this case is also analyzed below.

SYMMETRIC CHEMI-CRYSTALLIZATION IN A PARALLEL-SIDED MOLDING Prior to UV exposure, the volume of an element of material at a chosen depth within the molding is given by: [1 ]

where me,o and ma,o are respectively the masses of the crystal and the amorphous fractions before UV exposure, and Pe and Pa are the densities of the crystal and the amorphous phases, respectively. The initial crystallinity is given by:

f == c

m c,o m c,O+ma.o

[2]

The volume after UV exposure is given by: [3]

where me and rna are the masses of crystal fraction and amorphous fraction after UV exposure respectively. Therefore, if the change in crystallinity is ~fe, the crystallinity after UV exposure is given by:

[4]

143

Residual Stress Distribution

[5]

The corresponding volume change is: ~

V--

m C,O - m C

+ ma,O

- m a - (

m C,O - m C

)

[1 1J

[6]

From equations [5] and [6] it follows that [7]

The volume strain is given by:

)[_1 __ J== ~f P [_1 __1 J

tlfe(m e +m a 1 ______ P_c_P_a_ (me +m a )

[8]

P av where the average density is:

[9] If the strain is isotropic, the linear strain, E, can be written E

==!3V ~ V ==!3 ~fC[p a - P PcP a

J[f P C

+ (1 -

f)p ] == a~f C

[10]

It is the relative changes in strain through the depth of the molding that cause changes in residual stress. Uniform shrinkage would not cause any change in stress. Thus the average change in strain, Eo, must be subtracted from the strain, E, at each location to determine the change in stress there. Eo is found by solving the following equation:

144

Weathering of Plastics

S(c -CO)dZ1 where

=0=

S(a.1fc -c

)dz1

[11 ]

is the distance from the mid-plane of the molding and 2zo is the total thickness. It is next necessary to choose a function to represent the variation oft through the depth of the molding. This follows the variation of i1fc • The distribution of i1fc has been obtained from fractional crystallinity measurements made by Rabello on polypropylene samples exposed to UV for 9 weeks. 14 i1fc was quite significant near the surface, rising to about 0.04 at about 0.2 mm from the surface, then fell fairly sharply. The distribution ofi1fc is approximated by two 3.0 2.5 2.0 1.5 1.0 2z o straight lines to simplify the mathematical analysis (Figure 1). Using the parameters defined in Figure 1 it is found that

Figure 1. Fractional crystallinity changes within a 3 n1m thick PP bar after UV exposure.

a(B b +B b

o 1 1 2) £0 = - - - - - -

2z 0

[12]

For relatively thin bars (for example, 3 nun thick as used in the study described here) the strain given by [12] will actually occur in the thickness direction (z-direction). In the x-y plane the constraints are such that the strains are not permitted and instead an opposing stress must appear. Stresses in the x-y plane are expected to be equi-biaxial and the change in stress corresponding to a change in strain of (£ - £ 0) is -E(£ - £ 0)/( 1 - v) where v is Poisson's ratio. Thus the changes in residual stress, i1cri, through the depth of the bar, can now be obtained by substituting for i1fc • The changes in residual stress are given in terms of the distance from the surface, (zo - 21)' Thus, near the surface, ( 0 < 20 - ZI < b 1)

Residual Stress Distribution

145

[13]

In region b l < Zo

- ZI

< b2

[14]

In the central zone, ( b2 < Zo

- ZI

< 2zo - b2) [15]

Equations [13 ]-[ 15] can now be used to calculate the modifications to the residual stress magnitudes in the different zones. ONE-SIDED CHEMI-CRYSTALLIZATION IN A PARALLEL-SIDED MOLDING

When chemi-crystallization occurs non-symmetrically, the molding warps: the change in curvature must be taken into account when calculating the strain through the depth and when cOlnputing the equilibrium conditions. The strain at each depth is proportional to the distance from the Inid-plane of the molding, ZI: it is equal in magnitude to zl/R where R is the radius of curvature. In the following analysis the changes in fractional crystallinity at the exposed surface are taken to be similar to as those shown in Figure 1 whereas it is assulned that there is no change at the unexposed surface. A similar process to that given above for the symmetrical case gives the following for the average strain: Eo

a(B b1 +B1b 2 ) = = o- ---4z o

[16]

The curvature (==I/R) changes until Zo

-Zo

10" i,UV dz

[17]

146

Weathering of Plastics

where O\uv is the residual stress distribution after UV exposure. Thus if l/R is taken to be the change in curvature, this is given in tenus of the changes in strain by [18]

where E is Young's modulus. Substituting for £ and £0 gives: [19]

This result is important in that it predicts the curvature (warping) of the n10lding. It is also required in the calculation of the changes in residual stress, as given below: When 0 < Zo - ZI < b 1 [20]

[21 ]

when b2 <

Zo - ZI

< 2zo

[22]

EXAMPLES OF MEASUREMENTS OF RESIDUAL STRESS CHANGE SYMMETRIC CASE 14

It has been found by Rabello that the change in crystallinity in polypropylene bars exposed to a UV source in the laboratory was almost symmetrical even though the illumination was applied at one surface only. This is probably a consequence of two features of polypropylene

147

Residual Stress Distribution

pp N

2 ....

-~~

-...... Z

~ 0" <:2

"\ "-

-1

'-~ '\

,, "

}----~\l""""\_---~---/~--::O"'---1 \

-2 _.

-3

s,-moIJ'd~d

... ~ - strut

bl'.

e~an9f)

I'l()

¢3us"d by erysuHinHy etunqel

" "

.......

....

o,:,,",",.-,....-o~

_........

/' ~

-4

-s

-2 0

0.5

1.0

1.S

2.0

2.5

3.0

l.-.-l~~.........J.,,-""""...-i-.....-t.--'--.......l...........l-.-~"--oi--"~-.J

0.2

0.4

0.6

0.8

%0-'1

Figure 2. Residual stress distribution in an as-molded PP bar (solid line) and the change in stress caused by crystallinity changes due to 9 \veek UV exposure.

1.0

1.2

1.4

1.5

fmml

Figure 3. Residual stress distribution in PP bar after 8 weeks UV exposure (broken line) compared with the Slun of the profiles sho\vn in Figure 2. (analysis for half the bar only).

photo-degradation: firstly the UV transmission through polypropylene is reasonably great so that the UV level at the unexposed surface is still significant in a 3 mm bar; 15 and secondly, the process is mainly limited by diffusion rather than the illulnination level everywhere except very close to the surface. The depth-dependence of the change in fractional crystallinity after 9 weeks exposure derived from Rabello's measurements is given in Figure 1. The same study gave fc == 0.505 at the surface. For polypropylene, Young's modulus was taken to be E==I.3 GN/m 2 and Poisson's ratio v == 0.4; Pa is 850 kg/m 3 and Pc is 940 kg/m 3 .16 This gives

u==-033. Residual stress distributions have been measured using the layer removal procedure. 1-3 In as-molded PP bars the stress distribution is quasi-parabolic (Figure 2: solid line). After UV exposure the distribution alters quite significantly and an example measured after 8 weeks exposure is given in Figure 3. The change in the residual stress distribution through the depth of the bar was calculated froln equations [13 ]-[ 15] using the data listed in the previous paragraph and is presented in Figure 2 (broken line). If the two plots given in Figure 2 are added, the result should be the residual stress distribution after the exposure that caused the predicted changes. This sUlTIlnation is shown in Figure 3 as the "calculated result". Although it does not coincide with the predicted stress distribution the measured distribution is much closer to the modified ("calculated") distribution than to the as-molded distribution.

Weathering of Plastics

148

c--;-

e -... z

0 -I

...... ... .. -"

-2

C.l

0.4

0.6

U.S

to -Zt Figure 4. Fractional crystallinity changes within a 3 mm thick GFPP bar after 8 weeks exposure (notional).

1,0

1.2

1,.1

(mmJ

Figure 5. Residual stress distribution in an as-molded GFPP bar (solid line) and the change in stress caused by crystallinity changes due to 8 weeks UV exposure.

ONE-SIDED CHEMI-CRYSTALLIZATION In a study of the photo-oxidation ofglass fibre reinforced polypropylene (GFPP) it was found that the molecular weight degradation near to the unexposed surface was much less than that near to the exposed surface. 15 This was attributed to the very low levels of UV at the unexposed surface. It was found that the UV in adiation penetrated to a depth of approximately 0.8 nun from the exposed surface. No measurements are available for crystallinity changes with depth for GFPP but an analysis is attempted here in which, ~fc near to the exposed surface is taken to be the same as in unfilled PP but is taken to fall linearly to zero at Zo - ZI = b2 = 0.8 nun; ~fc is taken as zero for b 2 < Zo - ZI < 2zo . This is summarized in Figure 4. The value of a obtained from the definition given [10] was multiplied by 0.92, the volume fraction of PP in the GFPP grade used. The curvature (1/R) calculated from [19] is 0.331 m-I. This compares reasonably favorably with the measured value of 0.44 m- I for a bar exposed for 8 weeks. The distribution of residual stress change through the depth of the bar obtained using this value for 1/R and equations [20]-[22] is given in Figure 5. Figure 5 also shows the measured distribution of residual stress in an unexposed bar. The two curves in Figure 5 are added together to give the "calculated" result in Figure 6 and this result is compared with the measured distribution obtained from a GFPP bar exposed for 8 weeks. The agreement is very good. 4

149

Residual Stress Distribution

r;-

:k--~ ----I

I '-'-, . . r" -..... '. I' "\'

0[1"---- ,., ,; :'3'

::! -

e~'~u~.~~~

/ / -'

The modifications to

resi:U:~~tr~:::~r~

butions that are calculated from the changes in crystallinity caused by chemi-crystallization have been shown to account for a significant part of the changes that are actually

observed to occur. Exact agreement between the calculated and the measured dist tributions does not occur, indicating that -'-..I.-""---.J.-..l........- ' -__..:....~"'"___'"-oJ o Q.Z :)4 0& 0.8 1.0 12 l.4 1,(1 other effects are also present. The most likely source of discrepancy between the calculated and the measured distributions is Figure 6. Residual stress distribution near the exposed face of stress relaxation. It must also be admitted GFPP bar after 8 weeks UV exposure (broken line) compared that the ,1f profile for PP upon which the c with the sum of the profiles shown in Figure 5. symmetrical analysis was based is highly simplified and that the PP bars did not degrade symmetrically when illuminated on one surface only. It is interesting to note that the analysis for GFPP gave better agreement between the calculated and the measured residual stress distributions for exposed bars even though the chosen ,1fc profile not confirmed by measurement. Furthermore, there was evidence that some degradation occurred at the unexposed face 15 but this was ignored when constructing Figure 4. The reason why the agreement between the calculated and the measured results is better for GFPP than for PP may be that relaxation of residual stress is much more difficult when fibers are present. 5 _

mellSUfSC

,-I

CONCLUSIONS Photo-oxidation of crystallizing polymers may lead to chemi-crystallization which causes significant changes in the residual stress distribution. The major effect is to change the stresses near the surface from compressive in the as-molded state to tensile. This has been observed to occur in previous studies ofoutdoor weathering of polypropylene and is believed to contribute to the reduction in strength. Calculations have confirmed that a large fraction ofthe change in residual stress can be accounted for in terms of crystallinity changes but that there must be at least one other process (probably stress relaxation) occurring as well.

REFERENCES I.

B Haworth, C SHindle, G J Sandilands & J R White, Plast. Rubb.

PIVC.

Applics., 2, 59 (1982).

Weathering of Plastics

150

2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

J R White, Polymer Testing, 4, 165 (1984). A I Isayev & D L Crouthatnel, Polym. Plast. Technol. Eng., 22, 177 (1984). L D Coxon & j R White, Polym. Eng. Sci., 20, 230 (1980). C SHindle, J R White, D Dawson & K Thomas, Polym. Eng. Sci., 32, 157 (1992). M W A Paterson & j R White, 1. Mate1: Sci., 27 (1992) 6229. M W A Paterson &j R White, Polym. Eng. Sci., 33,1475 (1993). M M Qayyum & j R White, 1. Mater. Sci., 20, 2557 (1985). M M Qayyum & j R White, 1. Mate!: Sci., 21, 2391 (1986). M Thompson & j R White, POlylll. Eng. Sci., 24, 227 (1984). M M Qayyum &j R'White, 1. Appl. Polym. Sci., 43,129 (1991). J C M deBruijn, The failure behaviour of high density polyethylene with an embrittled surface layer due to weathering, Doctoral thesis, Delft University, Holland (1992). B Wunderlich, Macronlolecular Physics, Vol. 2. Crystal nucleation, growth, annealing, Acad. Press, New York, (1976). M S Rabello & J R White, to be published. B O'Donnell & J R White, Polym. Degrad. Stab., 44, 211 (1994). PH Geil, Polymer Single Crystals, JViley-Intersci., New York (1963).

Residual Stress Development in Marine Coatings Under Simulated Service Conditions

Gu Yan and J R White Universit)J ofNewcastle upon Tyne, UK

INTRODUCTION Polymer coatings are used extensively for corrosion protection of metals in marine environments. Solvent loss and, in the case of thermosets, the curing process, causes shrinkage of the coating. When it is applied to a stiff substrate the shrinkage in the plane of the coating is resisted and bi-axial tensile residual stresses form. If application of the coating is made at a temperature different from the subsequent service temperature then there will be further residual stresses that result fronl differential thermal expansion of the coating and substrate. The coating will always have a greater thermal expansion coefficient than the substrate so if the service temperature is less than the application temperature there will be a further increment of tensile residual stress from this source. The stresses may lead to failure of the coating by causing it to crack or to detach from the substrate (flaking, de-lamination, blistering, etc.). It is therefore important to have methods to measure the level of residual stress so that its contribution to failure may be assessed. If the substrate is thin and if the coating is applied to one side only then the tensile stress in the coating causes the coating-substrate combination to bend to restore tTIoment equilibrium, with the coating side becoming concave. The severity of the curvature depends on the level of stress. The most common experimental method of measuring the residual stress in coatings uses a measurement of the curvature, from which the stress can be computed provided that the thicknesses and Young's nl0duli of the two components are known. In the research described here this approach was used and the behavior of a theml0plastic coating and a thermoset coating were compared during solvent evaporation/curing. The curvature was measured using a strain gauge, permitting continuous monitoring. The method was further extended to examine the changes in residual stress when the coating was submerged in water and when it was removed again to dry out. The coatings formed the basis of an experi-

152

Weathering of Plastics

mental undercoat-top coat system of the kind commonly used in marine applications and measuren1ents were also made on bi-Iayers.

EXPERIMENTAL SAMPLE PREPARATION

The coatings used here were experimental materials designed for marine applications. The thermoplastic, coded "Anticorrosive A" (or "AA" from now on) was a single mixture containing a vinyl copolymer, solvent (xylene), tar pitch, pigments (57% solids by weight, 38% solids by volume). The thermoset was a two-component system, prepared just before application as with any common commercial two-pack epoxy, and was coded "Anticorrosive B" (or "AB"). AB also contained hydrocarbon resin and pigment and the solvent was Shellsol A. The coating thickness was computed from the mass of the coating after the solvent had disappeared and the density of the solid residue. The coatings were applied to thin steel shim substrates, chosen because the coatings were designed for marine applications on steel structures. Substrates 100, 150,200 and 250 J.lrn thick were used both to determine the best value and as a check on the reproducibility of the residual stress measurements. The shim was cut into coupons 150 mm x 25 mm. Surfaces were prepared for coating and for strain gauge attachment using emery papers # 120, #500 and #800 and finally cleaned using acetone or xylene. A strain gauge was attached to the side of the substrate that was to become the uncoated side using a cyanoacrylate adhesive. A microcrystalline wax coating (M-Coat W -1) was applied over the strain gauge and lead connections to water proof them and permit operation when submerged in water. The substrates were held flat on a magnetic table and the coating applied by hand brush. The coatings were suitable for spraying but spray equipment was not available that could be used in or near to the laboratory in which the subsequent measurements were conducted. The coatings were allowed to dry or cure in a room held at 30±1°C and the strain gauge signal was monitored continuously. Solvent evaporation was normally monitored for 14 days after which time the changes recorded in curvature or mass were minimal in AA. Thus samples for investigation of the effect of water immersion or for overcoating with AB were dried for 14 days before the next phase of the experinlent. WET1 DRY CYCLING

Coated substrates with strain gauges attached to the uncoated surface were placed in an empty tank in a room at 30±1°C then submerged in distilled water at 30°C, taking care not to disturb the strain gauge reading during filling.

Residual Stress Development

153

The strain gauge signal was monitored for 24 hours or 48 hours then the tank was emptied carefully. The strain gauge reading was monitored for a drying out period equal to the initial immersion period then the cycle was repeated.

TEMPERATURE CYCLING Tests were conducted with the saInples imn1ersed in water at 5°C then at 30°C using a 48 hour dwell time.

EVAPORATION KINETICS The solvent evaporation kinetics of the coatings were investigated by measuring the weight changes on specially prepared substrates without strain gauges attached. The coating was applied and the first weighing made as rapidly as possible in a Mettler AT analytical balance measuring to 100 Ilg. Readings were then taken every 10 seconds for the first 10 minutes then at increasingly long intervals. Measurements were continued for 14 days.

ABSORPTION/DESORPTION KINETICS The samples used for the study of evaporation kinetics were then used to investigate the absorption and desorption of water. During absorption, the samples were immersed in water and removed periodically for weighing. Each time they were removed the surface water was removed with blotting paper, they were weighed, then returned to the immersion tank as rapidly as possible. At the end of 48 hours they were removed fron1 the tank, the surface water removed and they were allowed to dry out in rOOln air, taking weighings periodically. After 48 hours of drying out the samples were re-immersed and the cycle repeated.

YOUNG'S MODULUS OF THE COATINGS To calculate the residual stress froln the curvature of the film plus substrate requires knowledge ofthe Young's modulus ofthe coatings. This was measured using tensile tests conducted on dog-bone shaped samples cut from free films ofAA and AB. Measurements were made on samples as follows: (a) as-prepared; (b) after immersion in water for 6 days at 30°C; (c) after immersion in water for 5 days then dried out for 1 day at 30°C; and (d) after immersion in water for 3 days then dried out for 3 days at 30°C. The Young's modulus was calculated from the small strain part of the load-defom1ation curve, which was fairly linear. For the combinations ofcoating and substrate thicknesses used in this work the Young's modulus is not very critical in the measurement of residual stress.

CALCULATION OF RESIDUAL STRESS The residual stress in the coating was calculated using an elastic analysis that assumed that the curvature was spherical, that is that the curvature transverse to the coupon axis was equal

154

Weathering of Plastics

1.5

0.5

o~.u..u..~~~~~~~~~~~~

100

150

time

200

250

300

350

400

(hours)

Figure 1. Development of residual stresses during solvent evaporation in AA coatings of different thickness.

50 100 150 200 250 300 350 400 450 500 5S0

time

(hours)

Figure 2. Development of residual stresses during solvent evaporation in AB coatings of different thickness.

to that measured along the coupon axis. The analysis was basically that described by Corcoran. 1,2

RESULTS RESIDUAL STRESSES DURING SOLVENT EVAPORATION Figures 1 and 2 show measurements of the residual stresses in coatings of different thickness for periods of 14 days or more. The stress in AA coatings of thicknesses ranging between 113 I-lm to 200 I-1n1 converges to a common value of approximately 1.2 MPa after about 12 days (Figure 1). The coatings with thicknesses below 150 I-1m display a stress maximum (of nearly 1.8 MPa for the thinnest) at short times «10 hours) before decaying to the common value. For coatings thicker than 150 I-lm the rate of approach to the final stress value is progressively slower as the thickness is increased. With AB the stress built up most rapidly in the thickest coatings but appeared to be approaching a constant value after 14 days whereas the stress in the thinnest coating (212 J-lm) was still climbing after 21 days (Figure 2). The mass loss measurements showed a three stage process. The first stage is free surface evaporation, followed by a mixed kinetics stage, and finally diffusion controlled evaporation. 2 Since solidification does not proceed uniformly across the coating sonle stress build up occurred due to solidification near the edges while the mass loss characteristic was still in the 2 2 first stage. This effect was greater in AA than in AB.

155

Residual Stress Development

1.0

0.8

0.5

ft; 0..

en en

....!401 v

0.0

-0.5

-1.0

drying out

v drying -1.5

O......-L.-.-I<.-...L.-....l.o.....a-.-*--'--.-"-"'---'l..-............J..-......I.-~

time

100

120

140

150

(hours)

Figure 3. Development ofresidual stress in an AA coating 165 /lm thick during wet/dry cycling at 30°C (48 h period).

OUl

~-...a..-""---a.--a..-.w.--.a--.: ........--""---a-.....A-...Aoo-~A..-...J

60 time

'00

120

140

160

(hours)

Figure 4. Developn1ent of residual stress in an AB coating 212 /lm thick during wet/dry cycling at 30°C (48 h period).

RESIDUAL STRESSES DURING WET/DRY CYCLING Figure 3 shows the variation in residual stress in AA during wet/dry cycling with a period of 48 hours (that is 24 hours water imn1ersion followed by 24 hours drying out in a room at 30 De). The stress increased rapidly to about 0.4 MPa during the first 2 hours of water immersion then increased much more slowly during the remainder of the first immersion period. Some reduction of stress was observed during the following drying out period. For subsequent cycles the changes were more modest, but for each complete cycle the increase obtained during immersion was greater than the drop obtained during drying out. It is notable that the stresses recorded are all tensile: if the major effect were swelling of the coating by the uptake of water then the stress would have been compressive and the curvatures would have been in the opposite sense. The stress observed in AB was also tensile immediately after immersion (Figure 4) but it quickly reversed to become compressive within half an hour (see reference 2 for a presentation of the results with an expanded time scale). After a cOlnpressive minimum of about 1 MPa the stress magnitude in this coating (212 fJ.m thick) reduced for the remainder of the period of water immersion. On removing the water from the tank a further incretnent of compressive stress was observed but this was quickly reversed and at the end of the first dry period the stress was only slightly compressive. Further wet/dry cycles gave a compressive increment during the wet period followed by a larger tensile increment on drying out so that the net effect was a drift towards tensile stresses (Figure 4). The amplitude of change observed during a cycle increased progressively.

Weathering of Plastics

156

0.4

f""

," (

,"p

n; Q..

0.3

b~-Iay~r

0.2

in water

0.1

drying out

tAB on AA)

Iv:

ti 0...

'9 V

-Q,O

--0,1 -0.2 -0.3 30'C

time

100

120

160

(hours)

Figure 5. Development of residual stress in a hi-layer coating consisting of256 ~m AB on top of 179 Jlm AA during wet/dry cycling at 30°C (48 h period).

-2

140

.... 0.4 0

/1'1 V)

«>

.. ~

immer sion at S·C

lmmersion at 30·e

-4 0

100

time

150

200

250

300

(hours)

Figure 6. Development of residual stress in an AA coating 252 Jlm thick during wet (5°C)/dry (room ten1perature) cycling (96 h period).

Results similar to those for AB are shown in Figure 5 for a bi-Iayer of 256 Jlm AB over 179 Jlm AA. The main difference between Figures 4 and 5 is the stress scale, which is expanded for Figure 5. Note that for the bi-Iayers it is assumed that the application ofthe AB top coat does not change the characteristics of the AA and that the change in curvature of the AB+AA+substrate combination is caused by stress changes in the top coat only. Changes in stress in AA caused by absorption of solvent from AB are ignored. Detailed differences occurred in the stresses observed for different coating thicknesses and, for bi-Iayers, different combinations of coating thicknesses. 2 Tensile stresses of nearly 2 MPa were observed during the drying out phase of the second and third cycle of an AB coating 293 f.lm thick. In bi-Iayers the stress after several cycles depended on the relative thickness ofthe two components and could be either tensile (generally when AA thickness was greater) or compressive (generally when AB thickness was greater).2

TEMPERATURE CYCLING Cooling samples to 5°C produced large tensile stresses which relaxed significantly during the cold dwell (Figures 6-8). In AB the stress reversed on returning to 30°C and the stress changes were repeated each temperature cycle (Figure 7). In the AA coating there was a progressive drift to higher (tensile) stresses (Figure 6). Bi-Iayers showed behavior closer to AB than to AA (Figure 8).

157

Residual Stress Development

immersion at S·C

immersion at

bi-Iayer

30·C

AS on AA

2 ~

c..

'1 1

$I)~

... 4)

r"IUOWfW

-2

-1 T

-4

~........L......~......l-.l.~~~~~~~

t....1-::......;.iI..w...J........ "(...............

100

time

150

200

250

300

(hours)

Figure 7. Development of residual stress in an AB coating 235 J1m thick during \vet (5°C)/dry (roon1 temperature) cycling (96 h period).

immersion at S*C

immersion at 30·C

-2 0

100

time

150

200

250

300

(hours)

Figure 8. Development of residual stress in bi-layer coating consisting of264 J1m of AB on top of 142 J1m of AA during wet (5°C)/dry (room temperature) cycling (96 h period).

DISCUSSION The residual stress development in AA coatings was similar to that observed by Cro1l 3A who also found that the residual stress in thermoplastic coatings reached an equilibrium value that was independent of the thickness and that the thickest coatings took the longest time to reach equilibrium. Tensile stresses form as the result of the volumetric shrinkage that accompanies the loss of solvent. During the early part of this process the coating is still fluid and stresses begin to form only when sufficient solvent has been lost for the coating to develop some energy elastic resistance to deformation. The time dependence of stress build up is determined by the diffusion of solvent through the coating and by the relaxation processes in the coating. The concentration profile will be dependent on the coating thickness and the relaxation rate will depend on the concentration. It is thus curious that the final stress level should be independent of coating thickness. The residual stresses in AB thermoset coatings were also tensile but showed greater scatter in magnitude and did not always approach a steady value even after 22 days. Crol15 also investigated thermoset coatings but used a solventless amine-cured epoxy. In his studies the coatings developed compressive stresses when thin «55 Jlm, thinner than any of the coatings investigated in the current work) and tensile stresses when in the range of thicknesses used here. Croll could not use solvent evaporation to explain stress development and he attributed the tensile stress to structural changes during the curing process. He SUlTIlised that compres-

158

Weathering of Plastics

sive stresses were caused by swelling due to water absorption (from the atmosphere). No attempt was made to control the humidity in the experiments repolted here and the small lack of consistency between different runs with AB coatings may have been caused by different contributions from this source. In the case of thermoset coatings the diffusion of solvent becomes progressively more difficult as the polymer network develops and release of solvent may proceed for an extended period of time. When AB was overcoated on top of a dry AA coating, solvent release from AB was not only into the air at the free surface but also into AA at the interface between the two coatings. Solvent entering AA will cause swelling giving an increment ofcompressive stress so that the overall build up of stress was much slower than for a similar AB coating applied direct to the substrate and the increment of stress due to the AB coating was much less than that obtained with an AB coating alone. 2 The behavior of the coatings when immersed in water and on subsequent drying out requires careful consideration. The initial tensile stress observed in AA coatings has not been explained with certainty. It is speculated that water may plasticize the coating, assisting the escape of residual solvent (or some other minor component). Subsequent changes in stress on dry/wet cycling are small but the sense of the changes are opposite to those which would be caused by water swelling during immersion and reversal of this effect during drying out. It is as if water has occupied the free volume and provided attractive forces to draw the molecules closer together. After water immersion the measured Young's n10dulus ofAA was higher than after solvent evaporation and it increased still further if allowed to dry out partially. This could be explained if water acted both to plasticize the polymer and to provide stronger intermolecular bonds and if the water participating in plasticization was less tightly bound (and more easily lost on drying out) than that providing intermolecular bonding. An initial increment of tensile stress was also observed in AB coatings on water immersion, possibly caused by a similar mechanism to that in AA. After about half an hour this effect reversed and subsequently for all phases of the wet/dry cycling the changes in stress were consistent with swelling by water (giving compression) with reversal during desorption of water. The overall drift in stress in the tensile direction could be due to further solvent evaporation (assisted by water plasticization of the coating). Broadly similar results were obtained by Negele and Funke 6 using a simpler epoxy coating. Of perhaps greatest interest here are the results obtained with AB coatings on top of AA coatings. The results are explainable qualitatively in terms of water diffusing through the AB coating and on into the AA coating during immersion and then this process reversing during drying out. The concentration gradients will be complex and will cause significant inertia in the time signature of the changes. As a result of the different stress responses of AA and AB coatings to water the sense ofstress in the bi-Iayer coatings depended on the relative thickness

Residual Stress Development

159

ofthe two layers, with smallest stresses occurring when their thicknesses were approximately equal. The largest stresses were obtained during the temperature cycling experiments. Differential thermal contraction is believed to be responsible for the generation oftensile stresses of the order of 4 MPa in AB coatings on immersion into water at 5°C. Partial relaxation of this stress then occurred and this caused the formation ofcompressive stress when the sample was restored to a higher temperature. The behavior of AA was basically similar but with a drift towards a permanent tensile stress. AB on top of AA showed behavior sinlilar to that of AB.

CONCLUSIONS The highest residual stresses observed in this study were caused by differential thermal contraction between coating and substrate. A temperature change silnilar to that between a dry dock in a warm climate and the open sea gave stresses of 4 MPa and more, a significant fraction of the failure strength. Other sources of residual stress are complex and are probably highly specific to the coating composition. When using bi-Iayered coatings the changes in stresses were moderated somewhat and it appears that a significant and beneficial reduction in the stress magnitude can be achieved by appropriate combination of thicknesses of the two layers.

ACKNOWLEDGMENTS The authors acknowledge Courtaulds Coatings for providing the materials used in this study and for the provision of a strain gauge signal conditioning unit. We are grateful to M Buhaenko for advice and for stitnulating discussions throughout the project.

REFERENCES 1. 2. 3. 4. 5. 6.

E M Corcoran, 1.Paint Technol., 41 (1969) 635. Van Gu, MPhil thesis, University of Newcastle upon Tyne (1997). S G Croll, 1. Coatings Techno!., 50 (638) (1978) 33. S G Croll, 1. Appl. PO(1'11l. Sci., 23 (1979) 847. S G Croll, 1. Coatings Technol., 51 (659) (1979) 49. 0 Negele and W Funke, Progl: Org. Coatings, 28 (1996) 285.

Balancing the Color and Physical Property Retention of Polyolefins Through the Use of High Performance Stabilizer Systems

M. J. Paterna, A. H. Wagner and S. B. Samuels C)Jtec Industries, Research & Developnlent, 1937 West Main Street, P.O. Box 60, Staniford, CT 06904-0060, USA

INTRODUCTION Polyolefin usage is growing in many n1arkets, including construction, farming, consun1er goods, toys and automotive parts. Unfortunately, polyolefin atiicles will degrade and undergo loss of physical properties and change in appearance unless adequately stabilized. UV stabilizers are added to inhibit degradation during outdoor exposure. To combat degradation during processing and fabrication, polyolefins usually contain phenolic antioxidants (AO), potent radical scavengers, and one or more hydroperoxide decomposing secondary antioxidants (thioesters, phosphites). Several factors must be balanced when designing a stabilization package for polyolefins. The package must be cost effective and must maintain part aesthetics on aging. In addition, the package must ensure that the resin will process well and that the fabricated part will meet its targeted service life in the intended application. Since stabilization packages typically contain several additive components, the potential interactions, chemical and functional, of the additives cannot be ignored. For example, the additives in a stabilization package may interact synergistically,1,2 as in the case of primary and secondary antioxidants. Negative interactions between additives are also possible, and when unanticipated, these can lead to premature product failure and legal liability. An example of adverse additive interactions is the reduction in color strength that occurs for certain combinations of pigments and hindered amine light stabilizers (HALS).3

162

Weathering of Plastics

When a stabilizer package is exposed to environmental agents (ultraviolet light, acid rain, gaseous byproducts of fuel combustion, smog), additional complex additive interactions are possible which may adversely affect the article's appearance or retention ofphysical properties. For example, upon exposure to exhaust gases (which contain a high concentration of NO x), resins containing certain hindered phenolic antioxidants; will discolor. This phenomenon, known commonly as "gas fading", can occur during warehouse storage prior to or after fabrication or at anytilne during the part's service life. Samuels et al. 4 studied the effect of exhaust fumes on a series of HALS and antioxidant packages. They found that exposure to exhaust fumes greatly increased the rate of discoloration ofmost HALS/AO packages. The rate ofdiscoloration upon NO x exposure was found to be primarily dependent on antioxidant structure, but the HALS can also influence the discoloration rate. In order to avoid gas fading, it is possible to use a very low pKa HALS, like HALS-l, with an antioxidant prone to gas fading since the rate of discoloration with this blend is very low. However, this combination will result in the sacrifice of physical property retention since HALS-l is not a high performance HALS. Formulations containing high performance HALS and a gas fade resistant antioxidant will not discolor upon NO x exposure. With care in formulating, it is possible to achieve excellent UV performance without encountering gas fade discoloration. An exan1ple is the combination ofHALS-2 and the gas fade resistant antioxidant, AO-l, a 1,3,5-tris(4-tert-butyl-3-hydroxy-2,6-dimethylbenzyl)-1 ,3,5-triazine2,4,6-(IH,3H,5H)-trione. The latter also affords excellent processing protection. 4 In many applications, adequate protection of the resin can be provided by a single UV stabilizer. In some systems, however, it is advantageous to use UV stabilizers with complementary mechanisms. UV absorbers competitively absorb the radiation, reducing the an10unt reaching the chromophores (impurities, microstructural features) in the polymer, thus reducing the photoinitiation rate. Hindered alnines are multifunctional as well, and will trap radicals and decompose peroxides at use temperatures. It has been suggested that HALS will also quench excited state complexes. Stabilizer packages containing antioxidants, HALS and an UV absorber are commonly used. The current study builds upon the previous work4 by investigating the interactions ofUV absorbers with antioxidants and HALS. To elucidate the relative contributions of the HALS, UV absorber, and antioxidant components to gas fade color formation, studies were conducted to determine the relative rate of color development in polypropylene (PP) formulations prepared with systematically varied HALS/UV absorber/antioxidant combinations. The formulations were exposed to the fumes of methane combustion. These results were compared with those results of additive saturated filter papers.

Color and Physical Property Retention

163

EXPERIMENTAL FILTER PAPER Five percent (wt/wt) solutions of the additives were prepared in methylene chloride. Volumes ofeach solution were mixed to achieve the correct additive ratios. Cellulose filter papers were allowed to soak in the blended solutions for five minutes before being allowed to air dry. Paper color was determined with a Macbeth Color Eye Colorimeter under Lab conditions with illuminate C, 2° observer, specular component excluded, and UV component included. Filter papers were exposed in a United States Testing Co. Atmospheric Fume Chamber (Model 8727) with custom temperature control. The charrlber was maintained between 57-60°C. The papers were exposed for a total of24 hours. PLAQUES Solid additives were weighed into polymer powder and dry blended for five minutes. The blended material was melt-n1ixed in a Haake torque rheometer base equipped with a 0.75 inch 25: 1 single screw extruder. The polymer was processed at 50 RPM and 220°C melt temperature. Plaques 2 x 2.5 x 0.100" were prepared by compression molding at 275°C on a PI-II Press. Plaques were exposed in an United States Testing Co. Atmospheric Fume Chamber (Model 8727). The chamber was maintained between 57-60°C. The plaques were exposed for a total of 48 hours. MATERIALS A variety of UV absorbers (Figure 1) and HALS (Figure 2) were tested with and without AO-2, a 1:2 blend of tetrakis[methylene (3,5-di-tet1-butyl-4-hydroxy-hydrocinnanlate)] methane and tris(2,4-di-tert-butylphenyl)phosphite. Three classes of UV absorbers were tested: hydroxytriazines, hydroxybenzophenones, and hydroxybenzotriazoles. Within the benzotriazole class, three different absorbers were evaluated (UVA-2, UVA-3, and UVA-4). Five different HALS were evaluated. The HALS varied in basicity with pKa values fronl 9.0 to 5.7. 4 Mantel's Profax 6501 unstabilized polypropylene was employed in this study.

RESULTS AND DISCUSSION FILTER PAPER SCREENING TESTS As part ofthis investigation, a rapid screening method was used to predict the relative propensity of different stabilizer components and additive mixtures for color formation. The test involves impregnating paper filters with solutions of the additives. The impregnated papers are

Weathering of Plastics

164

then exposed in a gas fade chamber and color development is monitored as a function of time over a 24 hour period. UVA·1 UVA·2 The advantages of this method UVA·3 include its simplicity, flexibility, speed, and cost when compared to evaluations involving resin extrusion. Although the filter paper method is not without its shortcomings,4 it does allow the rapid UVA·S UVA·4 determination ofwhether UV absorbFigure I. UV absorbers. ers are susceptible to gas fade discoloration, the effect of UV absorber structure on gas fade 1;\ MIXTURE HALSo1 HALs-4 01' HAl.$. uw.s.-s discoloration, and the effect ofHALS HALSo2 HAL~ on the gas fade discoloration of UV absorbers. .• Gt«lPJi The discoloration of filters impregnated with either UV absorbers HALS-3 HALS-6 or HALS alone was quite low, as ilR..Vanety(l( lustrated in Figure 3. Although their eJrooPl structures contain phenolic moieties, hydroxybenzophenone, hydroxybenzotriazole and hydroxytriazine UV absorbers do not gas fade at a signifiFigure 2. Hindered amine light stabilizers. cant rate. In contrast, filters containing a high pKa HALS with a UV absorber discolored to a greater extent (Figure 4). A low pKa HALS and a UV absorber combination did not discolor under these conditions (Figures 5). With a given UV absorber, HALS structure will determine the degree of gas fade in UV absorber/HALS blends. As illustrated in Figure 4, the UV absorber structure can influence gas fade in UV absorber/HALS blends. UVA-l, representative of the benzophenone class of UV absorbers, exhibited the greatest discoloration of all the UV absorbers tested. Within the benzotriazole family, substitution on the phenolic ring in the ortho position will decrease the degree of gas fade discoloration. When chlorine is located in the 5 position, the initial color is increased but the rate ofdiscoloration is similar to that for the 5-H-disubstituted benzotriazole, UVA-3. The

W---J..-,..Jr

R.V~';el)'O(

(\lIu~dH:

C'C~ac..."1J~,

~~. ~~

. =t-4,.

165

Color and Physical Property Retention

3On-------;::==~ .Oht Z S · . j . f - - - - - - - - - - - - - l m5 hr

.U~

j;: ..

20.u-------------=--'-l

30.,..,-;:===;------------, 110 hr

mShr 1 - - - - - - - .24 hr

.::; 20

15 J-I----.....,..

!l:

~... 10

;..

;;..

o UVA-4

UVA·2

UVA-3

UVA·!

UVA·S

Figure 3. NO, mediated discoloration ofUV absorbers.

UVA·4

UVA·3

UVA·l

UVA·5

Figure 4. Effect of HALS on NO, mediated discoloration of UV absorbers.

ImHALS.31

1k - - - - - - - - - .HALS-4 ~ "tl

0.07

j;:

0.06

0.05

0.04 0.03 0.02 0.01

o With AO·l

Without AO-l

Figure 5. Effect of HALS on NO, mediated discoloration of Figure 6. Rate (dYUdt) of NO, mediated discoloration for PP UVabsorbers. formulations containing HALS. 0.15% HALS, 0.04% AO-2, 0.05% CaSt, 0.08% TBPP, 100 mil.

triazine UV absorber/HALS blend discolored at a slightly slower rate than the disubstituted benzotriazoles.

PP PLAQUES As demonstrated in the previous study4 in LLDPE, in a PP resin matrix, HALS alone (without phenolic antioxidant) do not discolor appreciably upon exposure to NO x. As shown in Figure 6, the rate of discoloration is quite low and appears dependent on the substituent on the piperidinyl nitrogen. However, when an NO x sensitive antioxidant is introduced into the for-

Weathering of Plastics

166

•.05

0.05

U4 0.04

:;; 0.03

;:::;

....

~ ~

,:a:;

~ 0.63

....

0.&2

0.01

0.02

Qil

0.01

-••01 UVA·S

UVA·!

UVA·l

NoUVA

Figure 7. Rater (dYIIdt) ofNOx mediated discoloration for PP formulations containing HALSIUVAs. 0.15% HALS, 0.\% UVA, 0.05% CaSt, 0.08% TBPP, 100 mils.

0 Figure 8. Rate (dYIIdt) of NO x mediated discoloration of PP formulations containing UV absorber!AO blends. 0.04% AO, 0.1% UVA, 0.05% TBPP, 100 mils.

mulation, the rate of discoloration of HALS/AO blend is up to three times higher than that of the HALS alone. The rate is dependent on the HALS structure (substituent on the piperidinyl nitrogen, pKa) and the antioxidant structure. For HALS blends, like HALS-4, the rate of discoloration is deterIIAI..s..J mined by the more basic component in the blend. While HALS/AO blends can gas fade, Figure 9. Rate (dYIIdt) of NOx mediated discoloration of PP formulations containing UV absorber!AO blends. 0.15% blends of HALS with UV absorbers containHALS, 0.\ % UVA, 0.05% CaSt, 0.04% AO-2, 0.08% TBPP, ing phenolic moieties apparently gas fade at a 100 mils. very low rate in the resin matrix (Figure 7). Blends ofUV absorbers with antioxidants do not gas fade at an appreciable rate, and the rates for the various classes are within experimental error, as illustrated in Figure 8. Introduction of HALS to these UV absorberlAO blends significantly increases the rate ofdiscoloration (Figure 9). The rate of discoloration appears to correlate with substitution on the piperidinyl nitrogen, which also lowers the amine's pKa . The resulting discoloration can be attributed to the oxidation ofthe antioxidant by NO x, which can be catalyzed by the HALS. Of the three UV absorber classes tested, the hydroxybenzophenone class showed the greatest tendency to gas fade in the presence of an antioxidant prone to gas fading and a HALS. The rate here appears to be the sum ofthe rates for discoloration of the HALS/AO and O.t

~'VA.I.IIAI~WVA,ZI!lIlAI-SrovA'51

167

Color and Physical Property Retention

UV/AO blends. However, in the case of the hydroxybenzotriazoles and especially, the hydroxytriazines, the overall rate of discoloration was somewhat attenuated. Hydroxytriazines may be the most resistant to NO x discoloration because the pKa value for the hydroxy group is higher than that of a typical benzotriazole or benzophenone. In addition, the hydroxy group is lTIOre sterically hindered. In these experiments, HALS-3, a high performance HALS with a methyl substituent on nitrogen, exhibited lower discoloration rates than the HALS blend, HALS-4. The pKa for HALS-3, a hom*ogeneous tertiary amine, is lower than that ofHALS-4. HALS-3 will afford state of the art UV protection and is more resistant to negative interactions with other additives than a secondary an1ine or blend.

CONCLUSIONS In many outdoor applications, it is advantageous to use a UV absorber/HALS blend. When using a formulation containing a gas fade sensitive antioxidant, a high performance tertiary amine HALS, like HALS-3 would be the best complement. If a UV absorber is added to the formulation, utilization of a tertiary amine HALS will ensure that the color contribution by the UV absorber is minimized. In applications where color retention is critical, the best balance of properties will be afforded by a gas fade resistant antioxidant, like AO-l, a high performance tertiary amine, like HALS-3, along with a nondiscoloring triazine type UV absorber, UVA-5. The outstanding performance ofUVA-5 will be the subject ofa forthcoming paper.

REFERENCES I 2 3 4

Pospisil, Po(rmer Degradation and Stability, 39, 103 (1993). Pospisil, PO(l'mer Degradation and Stability, 40, 217 (1993). Eng and Nolan, Polymer Stab. and Mod. '97 Conf., Hilton Head, SC (1997). Samuels, Steel, Wagner, SPE RETEC, Houson TX (1995).

APPENDIX MATERIALS COMMERCIAL MANUFACTURED DESIGNATION NAME IRGANOX'B 1010 AO-I CYANOX E 1790 antioxiclant AO-2 TINUVIN 622 HALS-I CYASORB ~ UV-3346 light stabilizer HALS-2 CYASORB B CEC 3529 light stabilizer HALS-3 TINUVIN 783 HALS-4 CHIMASSORB~ 944 HALS-5

BY CIBA CORP. CYTEC IND. CIBA CORP. CYTEC IND. CYTEC IND. CIBA CORP. CIBA CORP

Weathering of Plastics

168

HALS-6 UVA-l UVA-2 UVA-3 UVA-4 UVA-5

CHlMASSORB E 119 CYASORB ~I UV-531 light absorber CYASORBE. UV-5411 light absorber CYASORB ID UV-2337 light absorber CYASORB lt UV-5357 light absorber CYASORB ID UV-1164 light absorber

ClBA CORP. CYTEC IND. CYTEC IND. CYTEC IND. CYTEC IND. CYTEC IND.

Activation Energies of Polymer Degradation

Samuel Ding, Michael T. K. Ling, Atul Khare and Lecon Woo Baxter Healthcare, Round Lake, IL 60073, USA

INTRODUCTION In the study of polymer degradation and durability, there is little reliable, predictive methodology that is universally valid over wide spans of temperature and titne. Many of the high temperature "accelerated" oven tests have been deemed unrealistic for different mechanisms were prevalent. In the mean time, for practical reasons, experimental tin1e spans of much longer than a year are extremely difficult to conduct. In the medical plastics industry, products are frequently sterilized by ionizing radiation, which severely depletes the antioxidant package. Yet to conform to many regulatory requirements, a scientifically based estimate of post-irradiation shelf-life must be provided. Thus a better understanding on the time and temperature influence on the material's performance is a necessity for product introduction. In this study we have examined the Arrhenius activation energy as a function of temperature for many polymer systems important in the medical industry. Data from oxidative induction time (OIT), accelerated oven aging, and real time ambient storage to up to 23 years will be used to determine the functional behavior and quantitative significance of the activation energy.

EXPERIMENTAL AND MATERIALS Technique used in this study includes ASTM D3895-92 isothermal OIT from Dupont 1090 thermal analyzer with 910 differential scanning calorimetry (DSC) cell. Forced convection air circulating ovens were used at various temperatures to assess long-term oven age shelf-life with sample embrittlement as end-points. Morphological studies were done using a Reichert FC4E cryo-ultramicrotome to prepare undistorted material blocks for SEM analysis. SEM analysis was done with the lEaL

Weathering of Plastics

170

35CF-SEM after sputter coating with palladium for surface conductivity. In addition, other available characterization data were incorporated into this study. The n1aterials studied consist ofpolypropylene (PP), high density polyethylene (HDPE), low density polyethylene (LDPE), EPDM and polyester thermoplastic elastomers. Gamma exposure at various doses was conducted in a laboratory gamma cell at dose rates of approximately 6-10 KGy/hr.

RESULTS AND DISCUSSION Both the OIT and chemoluminescence data support the general mechanism of degradation where the primary alkyl free radicals hv are propagated through atmospheric oxy~-Jllo<~. ~02 .. gen diffusing into the polymer via the for~H/~ mation of peroxy and hydroperoxy free o· """"'. radicals (Figure 1). The rate limiting steps '------OOH in this complex chain reaction scheme de· H termine the overall degradation rate. In this Figure 1. Oxidative kinetic chain reaction. regard, the action of stabilizer, such as phenolic antioxidants, chocks a section of the degradation loop by eliminating organic free radicals, or decomposing the hydroperoxides, and becomes sacrificially consumed in the process. The activation energy of the degradation rate, as expressed in the Arrhenius form, will be affected by factors such as polymer composition, stabilizer package, and polymer morphology. O2

...............,.................

~-.,.

GAMMA RADIATED POLYPROPYLENE Catastrophic failures have been reported during the PP shelf life storage period. Intense investigation has come to the following hypothesis, that, long lived free radicals trapped in the crystalline domains slowly migrate towards the crystalline/amorphous interface where they react with available oxygen to form peroxy and hydroperoxy radicals and initiate degradation near the interface. 1,2 When sufficient number of the tie molecules between crystallites were cut through this chain scission process, PP's elongation is reduced dramatically and catastrophic failures followed. To establish that long lived free radicals do playa significant role in the post irradiation PP degradation, a PP film sample was examined by electron paramagnetic resonance (EPR) spectroscopy. A distinct free radical spectrum was detected at room temperature about 6 months after irradiation. Incidentally, the strong EPR signal was completely eliminated when the sample was annealed in a vacuum oven at 90°C, a temperature much above the glass tran-

171

Activation Energies of Polymer Degradation

10 '

'0' ~

Oven / " 10'

/' OIT

10'

10 • 2.4

•

Ql;!'r,,)

10 •

L:' 10 •

()"~'''lrwl

H;so 60 KCal1 Mole

Oamma PP 00•• = 20 KOy Ra,. - 6 KGyl hr

10 ' Dlua hom Mehan (l-l l)l

10 '

p.8QO. Thermal C!1ar. of P6'imGl~ Matetlfl1$,

2.6

lY'

f- 10'

/ 2,5

•

0:\,'.: e

2.7

2,9

3.0

3,1

3.2

100011< Figure 2. Combined Arrhenius plot ofOIT and oven times for gamma irradiated PP,

10 ., 2.0

2.1

2.2

2.3

2,4

2.5

2.6

2.7

e Tutl Ed

2.8

2,9

3.0

lOOO/K Figure 3. Bell Labs OIr and oven aging data on LOPE cable compound.

sition for the amorphous phase for PP, and well into the alpha relaxation for the crystalline phase of PP. 3 In a separate study, a radiation grade PP's OIT under air flow conditions of 100 mll min was determined and the result compared with the same film sample (about 130 micron in thickness) after 20 KGy of gamma exposure at about 6 KGy/ hr dose rate. To access lower temperatures thermal stability, where the OIT detection becomes difficult, the gamma-exposed films were subjected to oven aging at 90 and 60°C and their failure times noted. When the OIT and oven failure times were plotted onto the Anhenius plot for the gamma inadiated samples, a continuous curve with diminishing slope toward lower temperatures emerges (Figure 2). This kind of continuity of functional behavior ofOIT data at higher temperatures and oven stability data closer to ambient could, at least in principle, produce long-term property prediction based on on data, provided that the rate of slope change can be determined separately. Further experiments along this line of reasoning are being carried out cunently to explore the boundary of validity for several polymer systems.

LOW DENSITY POLYETHYLENE This continuous curve behavior was very reminiscent of the data on crosslinked low-density polyethylene cable compounds studied with OIT, oxygen uptake, and oven aging experiments at the former Bell Telephone Laboratories,4 (Figure 3). When the high temperature results were extrapolated by the Anhenius equation to lower temperatures, grossly and physically impossible optimistic results were obtained. An examination of the activation energies indicated a more than fourfold difference between the high temperatures and near ambient (Figure 4). This observation prompted the Bell Lab. researchers cautioning against using the high temperature OIT for low temperature durability predictions. Nevertheless, by

172

Weathering of Plastics

10' r-~"--"""--""""--~--,

300

r;d." h)<.- (l .....~!,;."".".,· ..; I) N.~ , ..v..:~"""", ;;Ar·"-..... ·~~l'X (),,\';I;

dH{KJIV-ClC»

,-..

$M 6::Aff,';: ".')<.:n"ICL v~, "i

10 •

.-,10 '

200

loo

..c

e::

....

•

10 '

•

°'0'

100 0'

•

••

-..._-.J

10 . , ' - - _......._""'-

100

"/H:.~,(J

Q,l

150

200

250

T(C) Figure 4. LDPE cable activation energies.

1.6

2.0

2.2

2.4

2.8

lOOOIK Figure 5. EPDM rabber OIr and oven aging.

250 ,..,......,...,..............,.....-,-y--r-r-r-y....-..,,-,...-...,---

""" I

nO·, EJ>O,", 1" Cl\J)

~ .....,

200

recognizing the curved nature of the durability function, one can indeed achieve realistic predictions. EPDM RUBBER

A new class ofthennoplastic elastomers was created when olefinic polymers (polyethylene, polypropylene) are dynamically vulcanized with a crosslinkable elastomer such as 100 ethylene propylene diene rubber (EPDM). These so-called thennoplastic vulcanizes are Figure 6. EPDM activation energies. quite resistant to oxidation and studies have been available on their stability over long period oftime. 5 A general-purpose thennoplastic sample of 50 Shore D hardness was chosen for the OIT study. Published data from a long-term oven aging study for 50% strength reduction was plotted on the same graph for comparison (Figure 5). The activation energies at several temperatures are plotted in Figure 6. 100

POLYESTER ELASTOMER

Polyester thennoplastic elastomers (TPE) based on polybutyleneterephthalate (PBT) hard segment and tetramethylene ether (PTMO) soft segments constitute an important class of medical elastomers because of their wide property range, solvent bonding capability, oxidative stability and processing ease. OIT measurements conducted in air at higher temperatures

Activation Energies of Polymer Degradation

173

10 ' •••• 0

10'

...

• •

10 • •

10 I 10 • ~

10 •

«*

••

75KJ fMole

118 KJ f Mole

.•

•

il'.'

• 10·' ...............,..........

'Ni'l

c !<,>·J<'fhh.V~wN,.,..,tJ"l·W

167 KJ/MoJe

.................~..........~-~ 2,3 2.4 2,5 2,6 2,7 2.8 2.9 3.0

~~~

1.9 2.0

2.1

2,2

lOOO/K Figure 7. OIT of polyester thennoplastic elastomer.

Figure 8. PP surface "fibrils" at 60x.

again coincided with the oven aging data reported in the literature (Figure 7). The activation energies at several temperatures are also calculated.

PP SURFACE EMBRITTLEMENT

For an electron beam irradiated PP film sample undergone oven aging at 90°C, a curious phenomenon was observed. About 3 weeks into 20 40 60 oven aging, surface fibrils orthogonal to the exTime (Days) 90 C posed film edge surface became visually obFigure 9. PP crack depth in microns. servable. Under optical and electron microscopic observation, these fibrils are showed to be shallow, surface cracks (Figure 8). These cracks appeared to grow in number and their depth (measured in cross section by SEM) increases linearly as a function of time. The linear crack depth growth significant accelerated at approximately 25% ofthe film thickness (or about 50% of the film volume) (Figure 9). The activation energies determined from the rate of surface embrittlement are: Temperature, °C 30 70 90 Ea, kllmole 16 41 82 10

PP ACTIVATION ENERGY FROM OIT, OVEN AGING AND REAL TIME AGING Recently, in the authors' laboratories, several prototype and production polypropylene bottles, which have been stored under ambient conditions for up to 23 years, were discovered.

174

Weathering of Plastics

10 J

.,.,=

=: !

i-'

0 10

.-l

2.$

2.6

2.7

2.3

txt

$.1

',2

3,3

3.:5

10001K Figure 10. Rate of degraded layer fonnation.

Figure 11. Combined OIT, oven aging, and real time aging data for PP.

This "find" could allow the calibration of our long-term durability prediction methods. When the OIT of these products were determined, an excellent linear relationship with storage time, pointing to the zero OIT time of about 30 years. Hence, we can state, with reasonable assurance, that the durability of this particular grade of PP in the thin film fonn, under ambient storage, is about 30 years. When this data was combined with newly generated OIT and oven life data, plotted in the Arrhenius form, a continuous curve covering nearly 8 decades (100 million folds) of time was obtained (Figure 11). And when the local slopes were measured and converted to the activation energy at various tenlperatures, again, a concave curve resulted. Yet another significant observation on PP is that the activation energy from thermal aging processes when compared with the activation energies of the rate of brittle layer formation, near identical results are obtained. This apparent "self-similarity", or near identical activation energies at the same temperature exhibited for different degradation measurement parameters and different grades of polypropylene, could lead to much simplified modeling and understanding of the degradation and durability process. Efforts are currently underway to gather more supporting data on this self-similarity and the utility it could bring.

SUMMARY A broad based study on the kinetics of polymer degradation was conducted. The Arrhenius activation energy was used as the parameter to follow the rate dependence with temperature.

175 For most systems, a monotonic increasing trend with temperature was evident. This finding explains the frequent observation that kinetic parameters obtained at high temperatures often lead to grossly optimistic results at ambient. For a polypropylene CopolYlner systeln, combined data from OIT, oven aging, and real time storage of up to 23 years, yielded one of the most complete data sets covering over 8 decades oftinle. When the activation energies from thermal processes were compared with the rate of surface embrittlelnent, a striking self similarity, or near identical activation energies at the same temperature were evident. This observation could lead to broader applications and further understandings on the polymer degradation.

REFERENCES 1. 2. 3. 4. 5.

G. N. Foster, in Oxidation Inhibition in Organic Materials, Vo1.2, J. Pospisil, P. Klemchuk eds., CRC Press, Boca Raton (1989). L. Matisova-Rychla and 1. Rychly, in Polymer Durability, R. Clough, N. Billingham, and K. Gillen Eds, Am. Chem. Soc., Washington, DC (1996). L. Woo, J. Palomo, T.K. Ling, E. Chan, C. Sandford, Medical Plastics and Biomaterials, 3, (2), 36(1996). H. E. Bair, Thermal Characterization of Polymeric Materials, p. 869, E. Turi Ed., Academic Press, New York, (1981). N. R. Legge, G. Holden, and H. E. Schroeder, eds, Thermoplastic Elastomers, Hanser MacMillan, New York, (1987).

Failure Progression and Mechanisms of Irradiated Polypropylenes and Other Medical Polymers

L. Woo, Samuel Y. Ding, Atul Khare, and Michael T. K. Ling Baxter International, Round Lake, IL 60073, USA

INTRODUCTION Ionizing sterilization is gaining popularity in medical device and packaging industry because of its convenience and lower cost. Concerns over worker exposure to ethylene oxide and temperature constraint of medical materials for high temperature steam autoclaving have made radiation sterilization more preferable. The mode of sterilization is a consequence of the high energy electrons released from the interaction of the gamn1a ray photons or electron beam particles with the materials being sterilized. These high energy electrons in tum interact with the DNA sequences in the microbiological burdens, through permanently altering their chemical structures to render them innocuous. The high energy electrons, however, can also initiate ionization events in the material being sterilized. It can create peroxy and hydroperoxy free radicals in the presence ofoxygen, and start the degradation cascade. This could result in an unacceptable color formation, excessive pH shifts and high extractable. Furthermore, the degradation could also lead polypropylene (PP) to well publicized catastrophic failure during post radiation shelf life storage. From product quality and application viewpoints, it is thus highly desirable to develop a simple and very rapid system to characterize radiation sterilized packaging materials.

EXPERIMENTAL AND MATERIAL Technique used in this study includes ASTM D3895-92 isothermal oxidative induction time (OIT) from Dupont 1090 thermal analyzer with 910 differential scanning calorimetry (DSC) cell. OIT was conducted under air flow condition of 100 CC/min. Stabilizer concentration was determined by high perfonnance liquid chron1atography (HPLC) from established cali-

178

Weathering of Plastics

bration curves. Failure morphology was examined by lEOL 35CF-SEM after sputter coating with palladium for surface conductivity. The materials studied consist of various polypropylene CPP) with different stabilizer package and low density polyethylene (LDPE). Gamma exposure at various doses was conducted in a laboratory gamma cell at a dose rate of approximately 6 KGy/hr. Forced convection air circulating ovens were used at various temperatures to assess long-tenn oven aging shelf-life with sample embrittlement as end points.

RESULTS AND DISCUSSION DEGRADATION MECHANISM

Both the sterilizing action and the degradation caused by ionizing radiation are believed to result from Compton secondary electrons from the primary interaction event. The high energy gamma photons or accelerated electrons (from the e-Beam source) interact with the atoms in the material, creating a secondary high energy electron and a recoiling photons or electrons. These electrons can lead to a series ofsecondary ionization events in localized spurs. The cascade is propagated until all the excess energy above the ionization threshold is dissipated. Thus from a single incoming photon or electron, a binary tree configuration of secondary electrons is generated, and they are responsible for the bioburden kill and material degradation. Catastrophic failures have been reported during the PP shelf life storage period. Intense investigation has come to the following hypothesis: 1) Long lived free radicals trapped in the crystalline domains slowly migrate towards the crystalline/amorphous interface where they react with available oxygen to form peroxy and hydroperoxy radicals and initiate degradation near the interface. When sufficient number of the tie molecules between crystallites were cut through this chain scission process, PP elongation is reduced dramatically and catastrophic failures follows. 2) Oxygen from the surrounding environment will diffuse through the sample surface and react with free radicals in the polyn1er. Since the outer surface is n10re exposed to atmospherical oxygen, the extent of degradation near the surface is much greater than that of the interior. A brittle layer is then formed and has the same effect as forming sharp notches on the sample, creating stress concentrations, and reducing the break energy drastically. To confinn that long lived free radicals do playa significant role in the post irradiation PP degradation, a PP film sample was examined by electron paramagnetic resonance (EPR) spectroscopy at room temperature; the PP filn1 was about 6 months old after an irradiation dose of25 KGy dose at about 10 Kgy/hr rate. It is seen that the free radical mediated oxidative degradation continues in polypropylenes long after the irradiation event. Incidentally, the strong EPR signal was completely eliminated when the sample was annealed in a vacuum

179

Failure Progression and Mechanisms

25 y::: 5.8702e-r.2 + 71. 131x R"'2::: 0.980

o SO

oven at 90°C, a temperature much above the glass transition for the amorphous phase for PP, and well into the alpha relaxation for the crystalline phase of PP. OIT OF PP

Confirming what have been widely reported in the literature,l we also found the OIT at H-O various temperatures was an excellent linear function of active antioxidant content (Figure 1). The exceptionally linear response of OIT at multiple temperatures strongly indicates the potential of using this o 0.0 O~1 0.2 0.3 method as a simple (minimum sample prepHlen.olic Antioxidant % ( HPLC ) aration), and very rapid (within minutes), although non-specific assay for active Figure 1. PP OIT vs. antioxidant concentration. antioxidant determination. using Recent publications 2 chemolulninescence as the detecting scheme also yielded linear responses of the induction tilne with antioxidant concentration, but an inverse exponential dependence with initial hydroperoxide concentrations. Both the OIT and chemoluminescence data support the general mechanism of degradation where the primary alkyl free radicals are propagated through atmospheric oxygen diffusing into the polymer via the formation of peroxy and hydroperoxy free radicals. In this regard, the action of the phenolic antioxidant is mainly that of a hydrogen donor in eliminating organic free radicals and becomes sacrificially consumed in the process. Since antioxidants in PP reside primarily in the amorphous phase, their effectiveness to react preferentially with primary free radicals governs the overall post irradiation stability of the material. Accordingly, it is interesting to assess the correlation between OIT and radiation dose based on different suppliers. Figure 2 shows PP from three different suppliers, and their OII at 200°C at 0 KGy, 25 KGy and 50 KGy doses. It was clearly seen that supplier A formulation, where the OIT vanishes after only 25 KGy of dose, was not as stable, or effective toward gamma radiation as the other two formulations. On the other hand, although PP formulations from suppliers Band C had experienced sizable OIT reductions, significant fraction of antioxidant had remained to protect against further degradation. In a separate study, OIT ofa radiation grade PP was detennined before and after 20 KGy of gamma exposure at about 6 KGy/ hr dose rate with the same film of about 130 micron in o o

180

Weathering of Plastics

loS

20 0

":::l

'"....

,....., 15 ~

:2 0

":::l

<':l ....

":::l

...

(3 10

;;:

S N

Ir,

('oj

E '0

o f--Io....t - -........-

104

102

Dose;;;; 20 KGy Rate - 6 KGyl hr

lOt ........-+-+-J...........................0006---1I--+-..........I--i-04-I--+-+

2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 3.2

PP SUPPLIER Figure 2. Dose effect on various PPs.

Irr(IOOO/K) Figure 3. Combined Arrhenius plot and oven.

thickness. The data was plotted in the AlThenius fotTn in Figure 3. It is clearly seen that the gamma exposure has significantly reduced the on throughout the temperature range studied. In addition, the reduction factor appeared to be relatively constant over the entire temperature range. To assess stability at still lower temperatures, where the OIT detection becomes difficult, the gamma exposed films were subjected to oven aging at 90°C and 60°C. Their failure times were noted when film samples became brittle. Interestingly, when the oven failure data were plotted onto the AlThenius plot with the gamma samples, a continuous curve with diminishing slope toward lower temperatures emerges. This kind of continuity of functional behavior ofOlT data at higher temperatures and oven stability data closer to ambient could, at least in principle, produce long tetTn property prediction based on OIT data, provided that the rate of slope change can be detetTnined separately. Further experiments along this line of reasoning are being carried out cUlTently to explore the boundary of validity for several polymer systems. POST RADIATION AGING Since significant fraction of antioxidant is lost during ilTadiation, and possible alkyl peroxide and hydroperoxides are deposited in the polymer, it is interesting to monitor the remaining thennal oxidative stability as a function of storage time. The residual antioxidant has undoubtedly great influence over the shelf life stability of irradiated polymer devices. In addition, various polymers with entirely different oxidation mechanisms could degrade at different rates. Two polymer films were selected for this study. The first was a polypropylene

Failure Progression and Mechanisms

181

(PP) random copolymer of about 2 wt% ethylene content, and the other one was a low density polyethylene (LDPE) film. OIT data were first determined soon after the irradiation event. After several storage time periods of up to 10 months at ambient temperatures, the OIT's were determined as a function of storage shelf life. For the LDPE, it appears that for both o OIT/OIT(tO) radiation doses of 30.6 and 57.3 KGy, over about 10 months of storage time, the OIT re10-3 1--_ _..-+-_ _.........'"'--_ _... main nearly constant, subject only to the 1 10 1000 normal degradation in air. However, for the t Cd) 100 case of PP, significant time dependence is Figure 4. OIT of post irradiated PP. Data normalized to time seen, Figure 4. After 10 months only 6% of zero. original OIT is remained. The reason is the steady oxidative reaction leading to the constant loss of antioxidants in the polymer. This observation is, as expected, yet another evidence supporting the view that there are significant oxidation initiation reactions taking place long after the irradiation event. (Hence the model that long-lived free radicals migrating from crystalline domains toward the amorphous interface is supported.) PP POST IRRADIATION PHYSICAL PROPERTY DEGRADATION In our laboratories, we have conducted over years, real time studies on post irradiation material properties. 3 A radiation grade PP sample prepared both in a extruded film of about 250 micron thickness and injection molded tensile bars of about 3.18 mm in thickness were compared. The extruded film samples were first cut into testing strips of about 12.5 mm wide before irradiation. The samples have been subjected to 35 KGy of gamma radiation. Six months into aging at room environment, elongation retention for the tensile bars and films is about 50% and 15%, respectively, and it remains relatively constant for the 3 years duration of testing. The difference in long-term elongation retention while all other parameters were held constant forces one to focus on the sample thickness as a major factor. Surface cracks are observed for films held over long periods at ambient temperatures, indicative of a degradation progression initiated at the surface. PP SURFACE EMBRITTLEMENT PROGRESSION For an electron beam irradiated PP film sample undergoing oven aging at 90°C, a curious phenomenon was observed. About 3 weeks into oven aging, surface fibrils became visually ob-

182

Figure 5. PP film (99 KGy e-beam). Surface "fibrils". 3 weeks aging at 90°C.

Weathering of Plastics

Figure 6. PP film (99 KGy e-beam). Crack depth after 3 weeks at 90°C.

servable, Figure 5. Under optical and electron microscopic observation, these fi90C • brils are showed to be shallow, surface cracks, Figure 6. These cracks appeared to grow in number, and their depth (measured in cross section by SEM) increases linearly as a function of time. The linear crack depth growth significantly accelerated at approximately 25% of the film thickness, Figure 7. The surface initiated crack propagation, and Time (Days) the final acceleration are strong evidence of 20 40 60 oxygen initiated degradation with residual Figure 7. PP film (99 KGy e-beam. Crack depth after 3 weeks free radicals in the polymer bulk. At about at 90°C. 50% conversion, the surface area available for oxygen ingress into the bulk is no longer limited by the original surface. Additional surface area created by the numerous cracks became dominant for the gas transportation. The degradation over time, from the outer surface of the polypropylene sample conforms to the oxygen diffusion limited model first proposed by Gillen and Clough. 4 In this model, oxygen diffusing from the surrounding are consumed through the degradation reactions, setting up a concentration gradient across the thickness of the sample. As a result, the extent of degradation near the surface are much greater than that of the interior, frequently

Failure Progression and Mechanisms

183

leading to experimentally measurable profiles in density (degraded vs. undegraded polymer), infrared carbonyl absorbance, or density of serial sectioned specimens. To confirm the above hypothesis, a gamma irradiated (50 KGy) PP copolymer (about 2 wt% ethylene) tensile bar was examined after 15 years of ambient storage. A layer of surface embrittled polymer degradation product appears to cover the entire surface, while the core of the sample remained relatively ductile. The scanning Figure 8. 50 KGy gamma irradiated PP tensile bar. 15 years electron microscopy (SEM) morphology ofa ambient storage. high speed fracture plane cross section, Figure 8, indicated about 100 micron thickness for the totally degraded layer. For polypropylene, the above discussion and observation can lead to the the following mechanistic conclusion. With an amorphous glass transition (Tg) ofabout oDe, polypropylene is not a tough material at ambient temperature. This is true especially when PP is subjected to high strain rates, while the glass transition is shifted upwards by the activation energy. The ductile-brittle transition for PP has been determined by Rolando et al. 5 to be dependent on molecular weight and thermal history, but in general in the range between 0.1 and 10 sec· l at room temperature. When the oxidative degradation started to progress from the external surface, the brittle layer as a fraction of the total thickness steadily increases with time. The brittle layer has the same effect as forming increasingly sharper notches on the material. Takanno and Nielson6 had determined the energy to break for polypropylenes as a function of single notch depth (h) and radius (r) expressed as the stress intensity factor K, where K= I + (h/r)o.5. Break energy drastically reduces as the K factor increases from I to 2, which is a mild notch. Hence, when a sufficient brittle layer has built up on the surface, even a slow defonnation can have the same effect as a high strain rate event, catastrophic brittle failures becomes the inevitable outcome.

SUMMARY Post irradiated PP material undergoes significant degradation even at ambient temperature, and its rate depends on the copolymer content and the amount and type ofstabilizer system.

184

Weathering of Plastics

Post radiation failure progression starts from the oxygen initiated surface degradation with residual free radicals in the polymer. Over time, oxygen diffusing fronl the surrounding environment is consumed through the degradation reaction, setting up a concentration gradient across the thickness of the sample with most degradation on the outer surface. As degradation progresses, the surface becomes brittle and cracks fonn which act as notches. Once the stress intensity factor reaches a critical value, catastrophic failure can occur. Also, long-lived free radicals trapped in the crystalline domains can migrate slowly to the amorphous regions where they can react with available oxygen to form peroxy and hydroperoxy radicals and initiate degradation cascade near the crystalline amorphous interface. When a significant number of tie chains are broken, severe reduction in elongation results. An OIT method was found useful in determining the stability ofPP at high temperatures, and the data may be extrapolated to ambient temperature to predict the shelflife ofPP by following the curvature, non-Arrhenius behavior of Figure 3. This method, if supported by further data, may provide a useful method for shelf life prediction.

REFERENCES 1. 2. 3. 4. 5. 6.

G. N. Foster, Oxidation Inhibition in Organic Materials, Vo1.2, 1. Pospisil, P. Klemchuk eds., CRe Press, Boca Raton (1989). L. Matisova-Rychla and 1. Rychly, Polymer Durability, R. Clough, N. Billingham, and K. Gillen Eds, Am. Chem. Soc., Washington, DC (1996). C. Sandford, L Woo, Radiat. Phys. Chem., 31, No.3, 4-6, pp 671-678,1988. K. Gillen and R. Clough, Irradiation Effects on Polymers, D. W. Clegg and A. A. Collyer Eds. Elsevier Applied Science, New York, 1991. R. J. Rolando, W. Krueger and H. Morris, SPE Proceeding, ANTEC, 657,1986. M. Takanno and L. Nielsen, J. Appl. Polym. Sci., 20, 2193, 1976.

Chemical Assessment of Automotive Clearcoat Weathering

o. Carter III, John L. Gerlock and Cindy A. Smith

Che111ical and Ph)Jsical Sciences LaboratolY Ford Research Laboratoly, p.o. Box 2053, MD. 3083 Dearborn, MI48121-2953, USA

INTRODUCTION The top, unpigmented layer ofa modem automotive paint job is known as a clearcoat. It acts to protect the multiple under layers, while maintaining a beautiful appearance of the car. The selection process for materials for use as a clearcoat is an arduous process involving multiple tests. The desired performance includes resistance to chemical degradation in a photo-oxidative environment. Another performance requirement is the shielding of all under layers from destructive radiation, and this over time as well. We are developing rational measures of the oxidative transformations of a coating and of the UV protection it renders over a vehicle lifetime. To this end, we will describe photoacoustic and microspectroscopic technologies applied in the infrared, and UV as means to assay the changes in chemical composition and UV protection produced by the clearcoat in complete automotive paint systems. We must acknowledge that in making chemicaltneasurements, we do not assess the physical tolerance of a technology to accommodate these changes. Chemical change gauges signal the likelihood of a problem, but cannot detail the eventual impact of these changes on the life of the paint system. Measurement of mechanical properties 1 and film erosion rate are examples of parallel physical metrics which facilitate a complete assessment.

BACKGROUND The use ofchemically based tools to evaluate coating systems is the subject of a recent review by Bauer. 2 The role of resin and cross-linker type, the role and effectiveness of hindered amine light stabilizers (HALS) and ultraviolet light absorbers (UVAs) are examined. Substrate and coating under layers also affect the final system performance. This understanding

186

Weathering of Plastics

was achieved through the extensive use ofchemical analytical techniques. Such tools have included electron spin resonance spectroscopy,3 IR spectroscopy,4 NMR spectroscopy,5 UV spectroscopy,6 Raman microspectroscopy,7's and hydroperoxide titration. 9 Ideally, what is needed for evaluation purposes is a set of techniques that can follow chemical changes in individual layers of a full automotive finish system. Recently, time-of-flight secondary ion mass spectroscopy has been used to document the extent of photooxidation of multi-layer coatings. IO This powerful technique has advanced research in the field of coatings but is at present too expensive to operate as a routine tool for material selection. We propose variants from IR and UV spectroscopies that are accepted in the routine analysis environment. These spectroscopies, however, will be used in ways that are not routine. The focus in this report is on the characterization of the clearcoat layer as part of a full paint system as it weathers.

INFRARED SPECTROSCOPY One could in principle apply infrared microspectroscopy to examine thin sections ofa coating in cross section. 1I This method suffers from basic limitations due to the physics of light. The wavelengths of usual IR microscope access are 2.5 to 14 Jlm or the mid-infrared (4000 to 700 em-I). The clearcoat thickness approximates 40 f-lm and may lose half or more due to erosion during weathering. To be quantitative, all of the aperture dimensions, defining the region to be examined, should be more than three times the wavelength of the radiation used. This is needed to nlinimize the effects of diffraction at the optical element defining the sample to be examined. Thus, 10 f-lm sample areas are too small and so there are few sample segments that can be examined in a clearcoat layer. Also, small apertures limit the source energy that can be delivered from the source, pass through the spectrometer, then through a totally reflecting microscope, a sample and on to the detector. This limitation can be overcome using synchrotron sources,I2 but these are not set up for routine use. In the future, array detector based systems may assist in this mode of characterization but at the present this is not practical either. Thus, the surface capable infrared techniques remain. In this regime there is the attenuated total reflection (ATR) method and the photoacoustic detection method. The ATR optical arrangement is a well understood way of obtaining top layer spectra I3 that requires intimate contact with the sample surface and controlled pressure to get reproducible results. Experimental techniques exist to accommodate this requirement for new, smooth and somewhat compliant coating samples. As samples age and become rough and brittle, satisfactory contact between the ATR crystal and the sample becomes questionable and spectra become unreliable. We settled on PAS detection for this analysis problem. There are difficulties and limitations on this method too, about which more will be said. The basis behind PAS and the associated limitations are described in the literature 14 for uniform, solid samples. This will provide a starting point for further understanding of the infrared analysis.

Chemical Assessment

187

ULTRAVIOLET SPECTROSCOPY Assessing the UV protection capability ofthe clearcoat with weathering is accomplished with UV spectroscopy. Unlike the case for infrared radiation, the examination of30 Jlm and thinner layers with 0.4 to 0.3 Jlm radiation lends itself quite well to the use of nlicroscopes and microspectroscopic techniques. UV microspectroscopy of 10 fJl11 paint system cross sections is the method of choice. PAS-UV spectroscopy has also been evaluated as a technique to examine weathered samples. This method proved to be enticing, but experitnental and theoretical complexity and the changes in certain variables, such as thennal diffusivity, with weathering led us to drop this technique for weathered samples. However, PAS-UV has been used to elucidate the effects of manufacturing changes on the distribution ofUVAs in clearcoats. I5

EXPERIMENTAL DETAIL All PAS infrared data are obtained on a Mattson Cygnus 100 rapid scan FTIR system equipped with a water cooled source, a variable source aperture at 50% and an MTEC 200 cell. A hyperbolic turning mirror is used to bring the maximum intensity onto the PAS sample. A 65% carbon fill rubber plug, one cm in diameter, is the reference (described elsewhere I6 ). Samples are removed from dried steel test panels with an 11 mm punch and die set with the centering dimple ground off. All spectra are transfonned using 4000 data points and one order of zero filling to give a spectral resolution of 8 cm- I and a digital resolution of 4 cm- I • A scan velocity of3.6 kHz and an assumed coating thennal diffusivity of 1x1 0- 3 cm2/sec yields a thernlal diffusion length at 4000 cm- I of---6 J.lrn and at 2000 crn- I of---8 J.lrn. Taking the sampling depth to be two thennal lengths, we are sampling between 12 and 16 J.lm into the coating. Thus, there is ample opportunity for erosion during weathering. The data is collected, transformed, manipulated, stored and displayed using Mattson WinFirst@ software on aPC. The UV spectra of a coating systenl in cross section are obtained using a Carl Zeiss, Inc. Axiotron MPM 800 microscope/spectrometer system. Spectra are obtained at a spectral slit width of 5 run from 450 to 250 run. The salnple is masked by 5 fJl11 x 10 fJl11 spatial and 10 Jlm x 20 Jlm field apertures. A 40x immersion objective is used. The paint sample is removed from the steel substrate by bending the panel until the paint cracks off. The paint chip is sandwiched with an adhesive (Pliogrip 7779/100 from Ashland Chemical) between two pieces of clearcoat of known UVA concentration and absorbance. In this wayan extenlal spectral standard is included with each sample along with a check on cross section thickness. The sandwich is cut into 10 Jlm thickness in an RMS Rotary microtome using a diamond knife at ambient conditions. Samples are mounted on quartz slides with quartz cover slips and an

Weathering of Plastics

188

..,!

~GS ~

! ,

:Q.8 A

,;:.', ;1/\~'~'7:

.~.

•

;':.'<;~~~c~. y~

' _ _. . . . - ; .

-:0-

,OJ

•

~~;

~oo;;,·;2~~~~i;:~~~,._:u B

W.rriw,,*k(;.ij m

n~"'~'- l~

~~ H ·l'\,;:ttt~11l.tl

index matching fluid of 10: I water and glycerin is used as the immersion fluid. Control of the spectrometer is through a DOS based interface on a PC which also stores the data.

EXPERIMENTAL RESULTS

In Figure lA, Band C the micro-UV and PAS-IR spectra typical of what is used in the evaluation of a paint system is presented. This paint system was weathered 48 months without catastrophic failure. These data are representative of what is found for generally well performing systems. A comparison (}of Figure 1A and B indicate that -1000 3500 3000 1500 2000 1500 lOOO 500 the UV protection provided by Wavellumber. C the clearcoat has diminished and Figure I. Spectral data for a single paint system from a natural exposure study. has developed a gradient into and A & Bare UV spectra for the new and 4 year sample, respectively. C contains through the layer. This would rethe PAS-IR spectra of the samples with the top one being the 4 year aged sample sult in an increased possibility spectrum and the bottom one being the new sample spectrum. for photo-damage in the subsequent layers. The pigmentation causes the elevation in the base line. The last segment shown in the aged sample is likely a mixture of clear and base coats. The PAS-IR spectra of these two samples are presented in Figure IC. These results illustrate the accumulation ofphoto-oxidation products and the associated changes to the chemistry of the coating. Infrared spectral features in the aged sample are generally broader and less well defined. Many features have disappeared, while some are still evident and almost unchanged except for baseline elevation. This set of data also indicates that there is a loss of vCH, in the 3000 cm- I region, as the baseline rises under these features. This observation is universal for series of aged samples. Spectra obtained in cross section of aged samples using a Raman microscope have demonstrated that there is no gradient of vCH. 8 Since these surfaces also scatter light as they age, as

Chemical Assessment

189

documented by gloss readings, the sample surface can scatter IR radiation out of the PAS cell before it can penetrate the sample to be absorbed. This would result in reduced PAS intensity. We propose to account for this by normalization of the oxidation related intensity to the vCH intensity. Do these results correspond to transmission IR results obtained on c1earcoats 4000 1500 3000 2500 1000 1500 1000 500 alone? The data for this c1earcoat as an isoW;)venumbers lated sample, new and aged, is illustrated in Figure 2. Transmission IR data for the same clearcoat as in Figure 2. The transmission spectrum and the Figure 1. Top spectrum taken after 2500 hrs of accelerated PAS spectrum are not identical. The PAS-IR weathering in a natural like tester. spectrum, acquired on a rapid scan interferometer, will examine an ever thicker sample at longer and longer wavelength (smaller cm'I). The spectrum reports on a region about three time thicker at 400 cm-] than at 4000 em-i. In addition strong bands will become saturated in intensity when the sample has absorbed all of the radiation available to it. Both of these phenomena result in differences in the relative intensities. However, all of the features and changes seen in the isolated c1earcoat are reflected in the PAS-IR spectra (compare Figure Ie with Figure 2). The changes seen from 2000 cm- i and lower are quite specific to a particular coating technology and are often quite intense. Therefore, this part of a coatings spectrum is of most use to the formulator with knowledge of the components in the system. Our goal is to be as generic as possible. For this reason, we will focus on the 4000 to 2000 cm- i region. When examining isolated c1earcoats with transmission IR, the intensity in the vCH region, 3125 to 2750 cm- I is seen to diminish with weathering. Thus, we use this intensity change to track the change in film thickness with exposure time. With aging the oxidation products accumulate and the broad region under the vCH, from 3700 to 2200 cm- i, increases intensity. These products are associated with the entire range ofvOH and vNH containing species produced by oxygen insertion and bond breaking. Although, these groups have different molar extinctions, we treat them as single type and integrate the intensity from 3700 to 2200 cm-] as well as from 3125 to 2750 em-I, using these points to establish the baseline for integration. The ratio of oxidation products to vCH is established by subtracting the vCH area from the total region area and dividing the result by the vCH area, to normalize for the film thickness. The changes in this ratio determined by subtraction of the corresponding ratio results for the original material from that of an aged sample spectrum. Thus, we have derived a measure of the oxidation, !l[(vNH,vOH)area/vCHareahrans.,exposure, of the aged sample. With transmission

Weathering of Plastics

190

IR and UV of films, we can track L'1[(vNH,\OH)area/vCHareahrans.,exposure film loss and the loss rate ofUV absorption, which can T be expressed as the time to no protection. The lJ. S. results ofjust such a study for some European CIcan:.>lIl. R and North American commercial clearcoats; are reproduced in Figure 3. The results were used to rank these materials as likely candidates. But these are not full systems and full l~lut~.n •• system results do not always conespond Cltt\l\·~ closely to the isolated film results. With the development of the "L'1Trans.,Expos." term for It • films as a measure of oxidation product accu.;u • t,: 1·ju,. (lI'lI'Jlh ,11 llt,un I mulation as measured in transmission, a Ii! MII)y l'ltlI«firlfl similar procedure can be applied to the PAS-IR data on complete systems. As the Figure 3. The accumulation of photo-decomposition products portion of the sample examined is limited to ("~trans,5K hrsBoco,Boro"x4), film loss (JIm x4) and evaluation ofUV the top layer and since the spectrum is most protection in coatings after 5,000 hrs of accelerated aging as representative of the outer most portion measured in isolated films by transmission. thereof,14 the oxidation product accumulation will be highlighted even more than in the transmission experiments. Since the intensity observed in the region used to calculate "L'1" is not overwhelming, the problem of PAS saturation is minimal. Confounding the fundamental understanding ofthese data is the gradient nature ofthe photooxidation process. As illustrated in Figure 1C, the accumulation of oxidation products is apparent, but at present, there is no easily applied model for analysis of the distribution of products in non-uniform samples. However, the L'1[(vNH,\OH)arealvCH area ]PAS,Aus. for this system proceeds from 0.72 to 1.3 to 3.25 for 1,2 and 4 years exposure in Australia. "L'1 PAS" will have different values compared to "L'1Trans " but rankings made on the basis of"L'1 PAs " are similar and agree with field experience. There are other intriguing insights obtained by the combination ofthese two techniques. In Figure 4, the spectral data obtained for new urethane system and aged samples of two colors are presented. The coating in the new state has not undergone all of the isocyanate chemistry possible. After weathering, the samples have lost the isocyanate band and developed a "L'1 PAS,Fla.3y" of O. 73 for the red sample, but the silver sample has a"L'1 PAS ,Fla.3y" of 1.78. The evidence of a color dependent photooxidation is reflected in the UV results. The UVA in the red sample has developed a concentration gradient through half of the clem'coat. The silver sample, however, has lost considerably more of the UVA and the gradient includes the <~\<.J

191

Chemical Assessment

R 14

e I

~,'

(I \

S I n I

2 4000 3500 3000 2500 2000

1500

1000

500

Wavenumbers

.\ : O~ h

;'0 ~ it r

, i

: ~_,

\_, ----'. \ l

~_\

~ 0.3

,:;~~~~~t~~7j:: 1\ (~

Dq,t!ltmlJ,:-f\\;ils)

t\'\ - ..

rA'l'tlllnlt~f"ll~1

Figure 4. A presents the PAS-IR spectra for new (black) and 3 year aged samples of two different colors, red and silver, and the corresponding UV microspectroscopy data for the 3 year red, S, and silver, C, samples.

entire thickness of the layer. The under layers of the silver sample would be exposed to destructive radiation much sooner than with the red sample. Although only 3 year samples are available, these techniques would have indicated a color related problem earlier. In similar fashion, the effects of UVA loss of a different UVA can easily be seen in Figure 5. The absence ofthe UVA after 3 years is obvious from the comparison of the UV spectra Figures 5A and 5B. The "Li PAS ,Fla.3y" from the PAS-IR spectra, Figure 5D grows from 1.19 at 3 years to 2.18 at 5 years. Although the systems in Figures 4 and 5 are not highly dependent on the UVAs for stability, the under layers will be left susceptible to UV radiation after the additive is consumed.

192

Weathering of Plastics

- .-!-~-_ ... _..)r).~~:~;1 ~..~~d ,1' J ~

-,d A. ~'1-;i'i>,

t.,·

A ~

}OS ;

fO f ~ le4

::,:::::,,~2,:H';~;~~Df~~<~·J:2 .J.'~..

I~ INIIll);tJfl';o."'~'t

I .

P A

S I

n t

4000 3500

30()0

2500 2000

1500

1000

·500

Wavenumbers

Figure 5. UV spectra ofcross sections of new (A), 3 year Fla. aged (B) and 5 year Fla. aged (C) sample with oxanalide UVA and the PAS-IR spectra (D) of the new (black) 3 year (blue) and 5 year (red) samples ofa candidate clearcoat paint system.

CONCLUSIONS PAS-IR and UV microspectroscopy are useful tools for the chemical evaluation offull paint systems. The results are encouraging and we hope to see them accepted as coatings ranking and evaluation methods.

REFERENCES I. 2. 3. 4.

M. E. Nichols, C. A. Dart, C. A Smith, M. D. Thouless and E. R. Fischer, Polymer Deg. Stab., 60, 291 (1998). D. R. Bauer, J. Coating Tech., 66, 57(1994) 1. L. Gerlock and D. A Bauer, 1. Polymer Sci., Polymer Lell., 22,477 (1984) and subsequent papers listed in I. D. R. Bauer and L. M. Briggs, in Characterization of Highly Crosslinked Polymers, ACS Symposium Series No. 234, S. S. Labana and R. A. Dickie (Eds.), American Chemical Society, Washington D.C. p. 271 1984, and other papers listed in I. K. R. Carduner, D. R. Bauer, J. L. Gerlock, M. Rokosz and D. F. Mielewski, Polymer Deg. Stab., 35, 219 (1991) and other papers listed in I.

Chemical Assessment

6. 7. 8. 9. 10. 11.

12. 13. 14. 15. 16.

193

1. E. Pickett and 1. E. Moore, Polymer Prep., 34, No.2, 153 (1993) and other papers listed in 1. 1. L. Dupuie, W. H. Weber, D. 1. Scholl and 1. L. Gerlock, Po(vmer Deg. Stab., 57, 339 (1997). W. H. Weber and 1. T. Remillard, 7h Annual ESD Advanced Coatings Technology Conference and Exposition, Sept 28-29, 1998, Detroit, MI. D. F. Mielewski, D. R. Bauer and 1. L. Gerlock, Polymer Deg. Stab., 33, 93 (1991) and other papers listed in 1. 1. L. Gerlock, T. 1. Prater, S. L. Kaberline and 1. E. DeVries, Polymer Deg. Stab., 47, 405 (1995). A) P. B. Roush (Ed.), The Design, Sample Handling, and Application of Infrared l\Iicroscopes, ASTM Special Technical Publication 949, ASTM, Philadelphia, Pa., 1987; B) R. G. Messerschmidt and M. A. Hartco*ck, Eds., Infrared 1\1icrospectroscopy Theory and Practice, Marcel Dekkel; Inc, N.Y. 1988. D. L. Wetzel and R. O. Carter III, Fourier Transfonn Spectroscopy: II th International Conference, J. A. De Haseth (Ed., The American Institute of Physics, 567 (1998). N.1. Harrick, Internal Reflection Spectroscopy, John Wiley & 50115, N.Y. 1967. A. Rosencwaig, Photoacoustics and Photoacoustic Spectroscopy, Jolm Wiley & 50115, N.Y., 1980; R. M. Dittmar, R. A. Palmer and R. O. Carter III, Appl. Spectrosc. Rev., 29,171 (1994) .. R. O. Carter III, Opt. Eng., 36,326 (1997). R. O. Carter III, M. C. Paputa Peck, M. A. Samus and P. C. Killgoar, Jr., Appl. Spectros., 43,1350 (1989).

Effect of Aging on Mineral-Filled Nanocomposites

A. Va. Goldman, J. A. Montes, A. Barajas and G. Beall

NanocoJ; Inc. D. D. Eisenhour

A111erican Colloid, Co.

INTRODUCTION Polymer composite products are often exposed to aggressive service environments such as chemicals, high tenlperature, or ultraviolet (UV) radiation, which tnay cause deterioration of their properties and some changes in structure. When a composite material is in contact with the causes source ofdegradation, the damage may occur in the matrix, in the filler, or in the interfacial region l ,2 resulting in substantial property deterioration when compared to the unexposed material. Previous studies have shown that the presence of fillers may enhance or inhibit the rate of matrix degradation3 ,4,5 and can also interfere with the effectiveness of stabilizers 6 ,7 and crosslinking promoters. 8 A number of studies have been carried out on artificial and natural photodegradation of polymer composites. These studies have indicated that the influence of fillers on the extent of chemical photodegradation varies widely depending on the combination of polymer and filler. Narisawa and Kuriyama9 studied the natural and artificial photodegradation of fiber-reinforced engineering thermoplastics and reported that filled grades are more resistant to weathering, as evaluated by their reduction in mechanical properties. Casu and Gardette lO reported that the photooxidation rate ofpolybutyleneterephthalate (PBT) containing 30 wt% of glass fibers is about 70% of that for the unfilled polymer. This was attributed to the low penetration ofUV radiation into the sample. They also showed that the mechanisms of photolysis and photooxidation do not change when glass fiber is present. Although much work has been reported in the area of polymer degradation, very little addresses the degradation of filled theml0plastic, especially nanocomposites.

196

Weathering of Plastics

PBT is an important high performance engineering thermoplastic well known for its combination of desirable properties. These properties make it suitable for use in different applications where it is exposed to outdoor weathering. Recently, we employed PBT as a matrix for nanocomposites. 11 Nanocon1posites are a new class ofn1ineral filled plastics that contain particles from 1 to 100 run across and offer above-average mechanical properties, heat resistance, and gas-barrier properties. In terms of degradation nanocomposites with a PBT matrix have not been as well docun1ented as other aromatic polyesters. This paper contains infonnation on the degradation of PBT thermoplastic matrix and mineral-filled PBT nanocomposites (MFPBT) under artificial weathering conditions. This investigation is divided into two parts. Part I is devoted to structural studies on materials as pre-requisites for degradation and some changes in structure during long-term exposure. Part II is devoted to the discussion of relationships between structure and properties of the same aged materials.

MATERIALS AND METHODS SAMPLE PREPARATION Samples were prepared by compounding, extruding, and injection molding pellets using a Cincinnati Milacron Vista/Sentry Machine. 6.5 wt% clay (montmorillonite) was added to PBT powder and mechanically blended. The blend was extruded through a single screw @ 50 rpm with temperature zones of225, 230,230, and 220°C. The blend and the neat PBT samples were molded under identical conditions. Initially, materials were dried 3 hours at 120°C then molded under a cylinder temperature zone of 135°C. The injection pressures for booster, secondary, and back were 1200, 400, and 50 psi, respectively.

EXPOSURE PROCEDURE Degradation often results in poor surface appearance that frequently makes the product unserviceable long before any significant depreciation in mechanical properties takes place. Damaging forces include UV light, moisture, and ten1perature. (The shorter, more damaging UV wavelengths are filtered out during weather). Light sources that emit UV (shorter wavelengths) are most commonly used for acceleration. Sources that are a good match with sunlight, especially in UV region, produce slower degradation, but n1ay allow better correlation with outdoor results. Moisture and temperature are also important variables that need consideration. In the present study, the combined effect ofmoisture, temperature, and UV radiation was investigated. A Q-U-V Corporation Weathering Tester with horizontal option (Model # QUV-HO) was used for the exposure.

197

Effect of Aging

The exposure procedures were established in accordance with ASTM G53-88 and ASTM D 4329-92 standards and utilized a fluorescent UV, condensation type procedure. UVA-340 (UV-A) lamps were used to attain the best available simulation of sunlight in the shorter wavelength region. The test regimen was two cycles per day consisting each of an 8 hour UV exposure at 70°C and a 4 hour condensation at 50°C. Special care was taken to rotate san1ples and insure uniform exposure. Measurements were made either weekly or bi-weekly.

MECHANICAL TEST PROCEDURES Each data point represents the average of at least a five sample set. Tensile properties for neat PBT and nanocomposite samples were determined using ASTM D 638-93; the impact resistance of notched specimens was determined according to ASTM D 256-93 standard. Physical properties were determined by using the following instrumentation: Instron 4467, Dynamic Mechanical Analyzer DMA 7e, and Pendulum Impact Tester Model-CS137. Failure Analyses were also perforlned in this study. Results are presented in Tables 1-3. All details are discussed in Part II of this paper.

Table 1. % Strain at break vs. exposure time for pure PBT and nanocomposite, 6.5% silica/PBT Exposure time PDT, hrs 0.00 216.80 412.80 652.50 910.90 1215.6 1776.6 2019.2 2245.5 3018.5

0/0

strain at break PDT 100.00 82.26 17.70 17.50 4.25 16.39 2.57 1.84 1.82 2.17

Exposure tinle composite, hrs 0.00 216.80 412.80 652.50 910.90 1215.6 1720.7 2008.9 2505.2 3068.0

0/0

strain at break composite 8.03 6.64 6.46 6.86 6.04 7.20 6.90 7.22 5.58 5.82

STRUCTURAL STUDY-X-RAY DIFFRACTION Structural investigations of the aged samples focused on the degree of crystallinity and crystallite size. Murthy and Minor's12 method of subtracting the amorphous scattering from the crystalline scattering was used to determine the relative degree of crystallinity. The crystallite size (La) in A was calculated using Peak #3 (Figure 1) from the X-ray diffraction pattelTI and the following formula:

Weathering of Plastics

198

Table 2. Stress at maximum load vs. exposure time for pure PBT and nanocomposite, 6.5% silica/PBT Exposure time composite, hrs 0.00 216.80 412.80 652.50 910.90 1215.6 1720.7 2008.9

2245.5

Stress at max. load PBT, MPa 51.00 54.30 54.40 52.90 50.52 52.3 nla 35.3 34.5

2505.2

Stress at max. load composite, MPa 53.80 54.30 53.30 51.40 50.73 50.49 50.63 50.6 51.6

3018.5

32.5

3018.5

50.6

Exposure time PBT, hrs 0.00 216.80 412.80 652.50 910.90 1215.6 1776.6 2019.2

Table 3. Impact strength vs. exposure time for pure PBT and nanocomposite, 6.5°A. silica/PST

0.00 216.80 412.80 652.50 910.90 1215.6 1776.6

Impact strength PDT, JIm 44.1 41.8 32.7 32.6 32.5 33.8 32.9

Exposure time composite, hrs 0.00 216.80 412.80 652.50 910.90 1215.6 1723.2

2019.2 2245.5 3018.5

25.2 22.8 23.9

2008.9 2505.2 3018.5

Exposure time PBT, hrs

L = a

180A (1t~(28) a cos 8)

Impact strength composite, JIm 34.20 33.00 32.00 34.00 33.70 32.70 31.10 31.70 30.50 21.10

[1 ]

where ~(28)a is the full width at halfmaximun1 (FWHM), 28 is the peak position, and Ais the wavelength ofeu Ka (1.542 nm).13

Effect of Aging

199

X-ray diffraction scans were acquired using a Philips X'Pert MPD diffractometer equipped with a Kevex solid state detector. All data were collected from 1 to 40° 28 in parafocus geometry using Ni-filtered CuK a radiation at a step size of 0.02° 28 and a counting time of 0.5 second per step. Because only relative crystallinity was of interest, no attempt was made to correct the III data for incoherent scattering or for Lorentz and polarization factors. Using the software Peaksolve, by Galactic Industries,14 a series of Gaussian a curves was used to model the crystalline Figure I. Peak solving spectra for the X-ray difraction pattern of and amorphous scattering contributions to pure PST with no exposure amorphous halo (4) and crystalline PBT XRD spectra. A ratio of the combined peaks (1,2,3,5,6.7,8,9). area of crystalline peaks to the total area was used as an indicator of crystallinity. Peaks arising from crystalline regions were fitted to Lorentzian curves for comparison to the results obtained from Gaussian fits. Absolute crystallinities were ~ 10 to 15% higher when peaks were modeled by a Lorentzian function; however, the relative difference in crystallinity among samples was similar regardless of the peak shape chosen. Analyses of duplicate scans run on several different PBT samples yielded crystallinity values within ±1 %. Peak fitting was performed over a 28 range from 5 to 37°. A total of nine peaks were fitted to each PBT XRD pattern; eight corresponding to scattering from crystalline regions and one to amorphous scattering (Figure 1). All peaks varied in position, amplitude, and width with the following exception. In all cases, an upper allowable limit of2.5° 28 was set for the full-width at half-maximum (FWHM) for crystalline peaks. This width corresponds to coherent scattering from regions of -32 angstroms in size at a 28 angle of ~25° using Cu Ka radiation. 12 For the crystalline peaks, initial positions were set at 8.79, 15.7, 17.1, 20A, 23.3, 24.9,29.3, and 31.1 ° 28 (Figure 1). Since all ofthe PBT XRD patterns displayed well-defined maxima and minima, all fits converged to stable solutions without the need to restrict the range of allowable positions for individual peaks. The degree of crystallinity for samples of pure PBT at different UV exposure times was estimated from X-ray diffraction (XRD) scans following the procedures outlined by Murthy and Minor. 12 This procedure was modified so that a single Gaussian peak was used to account for amorphous scattering. Peaks corresponding to scattering from amorphous and crystalline regions were fitted to the XRD traces, and the degree of crystallinity was estimated from the 6W,r-.--------

•

200

Weathering of Plastics

relative scattering intensity from these two areas. The results obtained are in good agreement with those reported by Murthy and Minor. The data indicates that if only relative differences in polymer crystallinity are of interest, which is most often the case, good results can be obtained by using a single Gaussian peak to represent amorphous scattering. This eliminates the need to use "standards," which are typically ofunknown crystallinity, to define the exact profile of amorphous halos. POLISHED SAMPLES

Figure 2. XRD patterns of pure PBT at different times of exposure (A) irradiated surface and (B) non-irradiated surface. Total times of exposure in hours: (I) 0, (2) 216.8, (3) 412.8, (4) 525.5, (5) 910.9, (6) 1215.6 (7) 1712.0, (8) 2195.9, (9) 2245.5, (10) 3018.5

Several pure PBT and nanocomposite samples, at different times of total exposure, were polished to different thickness, in order to study the effect that polishing has on structure, degree ofcrystallinity, and crystallite size. The polishing was done with an Ecomet 6 - Variable Speed Grinder Polisher at a motor speed of80 RPM with water and a 400 grit surface.

RESULTS AND DISCUSSION The crystalline structure of the polymer has a very strong influence on its stability. Crystallinity changes during the course of degradation. In the initial stages of photodegradation chain scission prevails, reducing molecular weight. Shorter chains are more mobile and thus are able to crystallize more readily. It can be presumed that any change in structure or hindrance of the active centers by the orientation of molecules affects chemical processes, the extent of which causes a complete change of mechanisms for photochemical and photo-oxidative degradation. The difference between crystalline and amorphous areas results in different reactions in each area. In addition, the concentration ofcrystallizing polymer has an effect on the thickness of the crystallite (since new crystallization centers are created due to the polymer degradation, which can occur during extended heating).

Effect of Aging

201

The XRD patterns for pure PBT at different durations ofUV exposure were :yr;;,N~~:::::::r.~ plotted for the irradiated (Figure 2a) and .. - ............A" non-radiated (Figure 2b) surfaces to ~~B:=:~~~ show the differences in peaks and crystallinity. The X-ray diffraction patterns for the irradiated and non-radiated surfaces ofthe nanocomposite were plotted (Figures 3a-b) to show the difference in peaks and crystallinity in the clay peak I' , I r at approximately 8.8°[28]. In general " there are differences between XRD pat1.. • ... ·1·1 8 terns of the pure PBT and nanocomposite, which is the clay peak for the nanocomposite between 0 and 10°[28]. Other differences will be explained later. To determine the limitations on the penetration ofUV light for the pure PBT and nanocomposite samples (at a speci" •• jh~,.-.-~..,....,..~..,....~..,., ..............~~~....-l fied duration of UV exposure) an ,. .f extensive study was performed. This Figure 3. XRD patterns for the nanocomposites 6.5% silica/PBT at study included the depth of UV light different times of exposure (A) irradiated surface and (B) penetration and thickness of the skin non-irradiated surface. Total times of exposure in hours: (I) 0, (2) layer before the core was reached by pol216.8, (3) 412.8, (4) 525.5, (5) 910.9, (6) 1215.6 (7) 1712.0, (8) ishing different layers from the irradiated 2195.9, (9) 2245.5, (10) 3018.5 surface of the sample. The drastic change in the failure mode is known to be caused by the changes in properties ofa very thin layer ofsurface material-much too thin to affect the overall stiffness. An understanding of this phenomenon could be used to prevent brittle failure in weathered samples, and improve the weatherability of materials. At 652.5 hrs of total UV exposure, the composite's X-ray diffraction pattern, for the non-radiated surface was very similar to the 59 J..lm polished irradiated surface (Figure 4). This indicated that the UV penetration was limited to less than 59 J..lm in depth. The composite was studied at 5790 hrs total UV exposure and polished to depths of25, 51, and 68 J..lm to find the limit of the UV penetration on the composite. The main differences between the unpolished and polished surfaces of the composite sample were observed between I and 10° [28], where variations in peak shape and intensity occurred. The X-ray diffraction pattern of the

...

""r Uta

tU'

...

It"

202

composite's irradiated surface at 5790 hrs of exposure shows a higher intensity (about 90%), and is narrower at 6.8 0 [28], compared to the 51 and 68 /lm polished samples. However, the 6.8 0 [28] peaks of the 51 and 68/lm polished irradiated surface of the composite were identical when superimposed (Figure 0 5a). The 6.8 [28] peak ofthe 25/lm polished sample had a higher intensity (about 35%), more narrow than the 6.8 peaks arising from the 51 and 68 Ilffi polished irradiated surfaces. The peak at 6.8 0 [28] of the composite's non-radiated surface was very similar to the 25 /lm polished irradiated surface of the composite (Figure 5b). Therefore, we conclude that a maximum limit ofUV penetration is ~25/lm by observing that the clay peak at 6.8 0 [28] after 25/lm polishing is similar to the non-radiated surface. However, at polished layers less than 0 50 /lm, the clay peak at 6.8 [28] increased

Weathering of Plastics

Figure 5. Superimposed XRD pattems from 1_10° 29 for the nanocomposite, 6.5% silica/PST at 5790 hrs of total exposure: A: (I) 25 11m polished irradiated surface, (2) 51 11m polished irradiated surface, (3) 68 11m polished irradiated surface. S: (I) non-radiated surface, (2) 25 11m polished irradiated surface C: (I) unpolished irradiated surface, (2) 25 11m polished irradiated surface, (3) 51 11m polished irradiated surface.

Effect of Aging

203

until it reached a maximum sharpness at 0 f.1n1 (unpolished irradiated surface). For the nanocomposite, a plot was made of the unpolished irradiated surface, and the 25 and 51 f.lm polished irradiated surfaces (Figure 5c). This plot shows that the clay peak's intensity decreases as the depth of polishing increases. A plot of the 25, 51, and 68 ~lm polished irradiated surfaces of the nanocomposite was made by superposition of the XRD patterns (Figure 5a). This plot suggests that the core of the sample has been reached at about 51 f.lm since UV light did not change the XRD pattern of the clay peak. Finally, from the plot of the XRD patterns for the nanocomposite's non-radiated and irradiated 25 f.lm polished surfaces (Figure 5b), we concluded that the depth of light penetration must be approximately 25 f.1m since the irradiated surface XRD pattern matched the 25 ~lm polished XRD pattern. The filler serves as a screen to block the UV light on nanocomposites and therefore itnproves weatherability. The PBT matrix sample at 5790 hrs total UV exposure was polished to 23,49, and 60 f.1m from the irradiated surface. First, the unpolished PBT sample from irradiated and non-radiated surfaces were compared at 5790 hrs of exposure. The XRD pattern for the unpolished PBT san1ple at 5790 hrs exposure of the irradiated surface showed a slight shift to a lower angle, and a decrease in intensity in the range 1-25° [28] compared to the non-radiated surface. However, the peaks for the irradiated PBT surface were sharper and the degree of crystallinity was higher (31.7%) compared to the degree of crystallinity of the unpolished non-radiated surface (28.2%). This effect indicated that for this particular duration ofUV exposure the degree of crystallinity of a sample increases for the UV irradiated surface. The crystallite size for the radiated and the irradiated surfaces did not differ significantly at 5790 hrs of total exposure. The X-ray diffraction pattern for the 23 f.lm polished PBT sample at 5790 hrs ofexposure showed a slight left shift and lower intensity from 1-25° [28] compared with the unpolished irradiated surface. This difference between the 23 ~lm polished and unpolished non-radiated surfaces correlates well with the results found for the unpolished irradiated and non-radiated surfaces. However, at 25° [28], the 23 f.lm polished PBT sample's X-ray diffraction pattern shows a greater intensity than that from the unpolished non-radiated surface. The greater intensity suggests a transition state ofthe polished sample's PBT structure. The peaks for the 23 f.lm polished samples were sharper and the degree of crystallinity was higher (31.3%) compared to the degree of crystallinity of the unpolished irradiated surface (28.20/0). The crystallite size of the 23 f.1m polished PBT sample (114.6 A) was not significantly different from the unpolished irradiated surface (110.4 A). The crystallinity and crystallite size of the 23 f.1n1 polished sample and the unpolished irradiated surface of the PBT sample at 5790 hrs exposure did not differ considerably. The XRD pattern for the 49 f.1m polished PBT sample at 5790 hrs of exposure was consistent with the other results in regards to the slight left shift and lower intensity from 1-25° [2

204

Weathering of Plastics

8] compared with the unpolished irradiated surface. However, the 49 Jlrn polished PBT sample showed a stronger intensity at 17 and 23° [28] compared with the unpolished irradiated surface. The degree of crystallinity of the 49 Jlm polished PBT sample (30.1 %) was slightly lower than the 23 Jlm polished sample (31.3%) and unpolished irradiated surface (31.7%), but still higher than the non-radiated surface (28.2%) of the sample. The crystallite size of the 49 flm polished PBT sample (124.6 A) was larger than the 23 Jlm polished sample (114.6 A), unpolished irradiated (110.2 A), and non-radiated surfaces (110.4 A). These results indicate that even after polishing, the degree ofcrystallinity ofthe irradiated surface did not drop below the degree of crystallinity of the non-radiated surface. The 60 J.Ull polished PBT sample's XRD pattern at 5790 hrs of exposure was in concordance with the other results in regards to the slight left shift and lower intensity from 1-25° [28] compared with the unpolished non-radiated surface. However, the 60 Jlm polished PBT sample showed a stronger intensity at 16.5 and 24.5° [28] compared with the unpolished non-radiated surface. The crystallinity of the 60 Jln1 polished PBT san1ple (31.0%) was not significantly different than the 23 Jlm polished sample (31.3 %) and unpolished irradiated surface (31.74%), but it was higher than the 49 J1m polished sample (30.1 0/0) and non-radiated surface (28.20/0) of the sample. The crystallite size of the 60 Jlm polished PBT sample (121.5 A) is higher than the 23 Jln1 polished sample (114.6 A), unpolished irradiated (110.2 A), and unpolished nonradiated surfaces (1 10.4 A), but it was lower than the 49 Jlm polished sample (124.6 A). There seems to be a transition at 49 Jlm of the polished inoadiated surface of the PBT sample. It was concluded that the effect ofthe UV irradiation on the samples was limited to less than 50 flm penetration for this study at these exposure times, where the crystallinity was higher and the crystallite size was lower. Our results correlate well with Casu and Gardette'slO results on photolysis and photooxidation. They found that the photooxidation and photolysis ofPBT were limited to the first 50 Jlrn superficial layers of the irradiated surface. The non-radiated surface of the PBT sample at 5790 hrs of exposure was polished to 43 Jlrn to compare the effect of polishing on the crystallite size and degree of crystallinity. The degree of crystallinity, 28.2%, was unchanged, however, the crystallite size increased from 110.4 A to 121.8 A. From this observation the core crystallite size was found to increase on both the irradiated and non-radiated surfaces. There was no significant effect on the degree of crystallinity of the surfaces. However, the degree of crystallinity of the irradiated surface did not change substantially. The effect of polishing on crystallite size may have been from concavity of the sample and differences between the surface and core of the sample injection molding preparation. Garbauskas and co-authors 15 described how different depths of an injection molded polymer represented different phase compositions. For instance, they mentioned that the outennost skin of the sample «0.01 flm) contained a mixture of

205

Effect of Aging

Table 4. Degree of crystallinity and crystallite size vs. exposure time for pure PBT irradiated surface Exposure time PDT, hrs

cos8 (deg)

Degree of crystallinity, 0/0

Crystallite size,

Full-\vidth at half-maximum, de2

0.0

26.38

95.36

0.94

216.8

29.11

100.36

0.89

412.8

29.69

108.56

0.82

625.5

29.68

107.54

0.83

910.9 1215.6

30.63 28.87

102.56

0.87

108.53

0.82

1712.0

30.87

100.13

0.89

2195.9

30.48

105.59

0.85

2245.5 3018.5

29.46 30.73

97.94 92.10

0.91 0.97

0.99

Table 5. Degree of crystallinity and crystallite size vs. exposure time for pure PST non-radiated surface Exposure time PDT, hrs

cos8 (deg)

Degree of crystallinity, 0/0

Crystallite size,

Full-\vidth at half-maximum, deg

0.0

26.38

95.36

216.8

27.21

99.55

0.90

412.8

27.87

111.78

0.80

625.5

27.44

107.20

0.83

910.9

28.42

104.72

0.85

1215.6 1712.0

27.32 28.22

109.68

0.81 0.91

98.15

0.94

2195.9

27.48

99.27

0.90

2245.5

28.08

97.75

0.91

3018.5

28.19

99.37

0.90

0.99

semi-crystalline and amorphous material. A layer of polish greater than 20 Jlm represented the "core" or bulk of the sample, with a transition between the skin and the core.

206

Figure 6. Peak solving spectra for the XRD pattern of pure drawn non-radiated surface of PBT at 216.8 hrs of total exposure. Crystalline sizes in A: (1) 31.32, (2) 54.11, (3) 36.27, (4) 28.38 (average crystallite size 37.52).

Weathering of Plastics

The investigation performed for drawn aged PBT samples and substantial differences in their structures and mechanical properties were found. The undrawn specimen was highly degraded and easily broke when touched. These observations suggest that cracking is mostly a surface-related phenomenon that occurs in amorphous regions. Photodegradation leads to chain scission, probably occurring in tie molecules (connecting neighboring lamellae), which permits the movement of polymer chains and allows retraction to occur after a sufficient amount of chain scission has occurred. Degradation ofundrawn PBT samples oc-

curs mainly by the Norrish Type II mechanism as opposed to the Norrish Type I mechanism 16 in drawn PBT samples. This difference is due to the orientation inhibition on the ability of polymer chains to twist into the conformation necessary for the Type II scission (Figure 6). The effects ofUV radiation and artificial weathering exposure on crystallinity, crystallite size, and elongation (Tables 1, 4, 5) were simultaneously investigated. It was found that artificial weathering resulted in an increase in crystallinity of specimens for both the irradiated and non-radiated surfaces (Figures 7, 8). This result correlates with Gooden's17 study of photochemistry and structure of poly(ethylene-co-(carbon monoxide)). After 200 hrs of total exposure, drawn specimens exhibited cracks, but retained at least part oftheir initial mechanical properties. From 0 to 412.8 UV exposure, the crystallinity rose to 29.7% and the crystallite size rose to 108.6 A, while the ultimate elongation dropped drastically. No significant changes were seen for the degree of crystallinity (29.7-30.9%) and crystallite size (108.6-100.1 A) until 1712 hrs of UV exposure for PBT. At 1712 hrs of UV exposure the crystallinity rose to its highest level of30.7%. The ultimate elongation dropped a second time to approximately 2.6%, while the crystallite size became lower, 100.1 A. Finally, at 3018.5 hrs and 5790.8 hrs ofUV exposure, the plot of degree ofcrystallinity versus time of exposure reached a plateau, but the crystallite size dropped to 92.1 A (Figure 7). A similar plot is shown for the degree of crystallinity and crystallite size versus time of exposure for the non-radiated surface (Figure 8). If the degree of crystallinity versus time of exposure for the irradiated and non-radiated surfaces are compared, the degree of crystallinity for the in adiated surface rose 4

Effect of Aging

207

,. r--....,.--~--r----r~-~-..-.,---r----. ". '"

'<+--:'~.-""""l:+-I""""'~"'+--+--+--t-;o''''''-+----t...

. Figure 7. Degree of crystallinity (I) and crystallite size (2) versus exposure time for UV exposed pure PBT samples on the irradiated surface,

: __J \,

:/\"

ll \ \ j . "IX: \ / f'·.· \ X

r-1..\

"'-:,.1

~I__

t"l_ _

more than for the non-radiated surface. This is because the irradiated surface was exposed to an alternating combination of condensation at (50°C) and UV light at 70°C, while the non-radiated surface did not have the effect of the direct UV light. Since the thickness of the PBT samples was approximately 3.2 mm, UV light did not penetrate to the opposite surface.

expo~~:e:he~~;sta~~tl::I~;

weight decreased and the ultimate elongation dropped (I-....,;f,f-.t--'l..t-+-",~-\j-/-r-':-\-+--+-..>r--+--+---I '.. due to defects that generated i" If 1 \1 ,.. 1 in the sample. Therefore, ~ embrittlement of matrix can " 1/,/ \_~.~~~-\. ~_._-- -_... .. be controlled by two associ1f/+--+----t---1,ir---t .. ated processes: reduction of molecular weight and in1, .1-, '"':-_~ ...~_"",_'--_~_l-_~_-l-. __..J. ...._ _...I_'--_....I_.. creased crystallinity.18 As a 'Ii

/,,1

...

result of morphological changes in the material's structure, chemical mechanisms are also changed. The degradation of the specimen leads to the formation of monoclinic forms and a reduction in crystallite size (Table 4 and 5). Morphological changes affect the diffusivity as well as other physical properties. This detailed study on nanocomposite and pure PBT shows that total crystallinity is not the only parameter that should be considered when interpreting structural data. Crystallite size, the type of crystalline structure, and their effect on conformation also need to be considered. Polymer morphology may, in some cases, be a greater influence on degradation than its chemical structure. 19 This study also shows that during the aging process some reorganization in structure occurs. The crystallite size of the PBT matrix first rose, then remained approxiFigure 8. Degree of crystallinity (I) and crystallite size (2) versus exposure time for UV exposed pure PBT samples on the non-radiated surface.

208

Weathering of Plastics

mately constant, and finally dropped. For PBT matrix, the crystallite size decreased starting at 1200 hrs of exposure and continued until 1700 hrs of exposure. At the same time, mechanical properties decreased. It is suggested that decreasing crystallite size is a criteria of long-term degradation and full reorganization of structure. We can also predict that if the crystallite size decreases mechanical properties will show a corresponding decrease (strain at break, maximum stress, impact strength).

CONCLUSIONS 1. 2. 3.

4. 5.

It was found that the depth of UV light penetration was limited to approximately 25 IJm for the nanocomposite and to 50 IJm for pure PBT. The filler serves as a screen to block the penetration ofUV light into nanocomposites and therefore improves weatherability. Chemi-crystallization probably occurs in the presence of filler by mechanisms similar to those that occur in the pure polymer, involving sholt-range motions of the molecules. UV radiation increases the degree of crystallinity of exposed surfaces. The degree ofcrystallinity is not the only parameter that should be considered for the reorganization of structure. Clystallite size and type of crystal structure are also important. During the aging process, some reorganization in structure occurs. The crystallite size of the PBT matrix initially rose, remained constant, and finally dropped. We suggest that decreasing crystallite size is an important criteria of long-tenn degradation of PBT matrix. Our results indicate that the service life ofthe polymer can be significantly increased by using clay (montmorillonite) as a filler.

REFERENCES 1. 2. 3.

4. 5. 6. 7. 8. 9. 10. 11.

Bank, L.C., Gentry, T.R., Barkatt, A., and Reinf, 1., Plastic Composites, 14, 559 (1995). Morgan, R., American Society of Composites- Second Technical Conference, Delaware, 250 (1987). Toennaelae, P., Suokas, E., Paeaekkoenen, E., and laervelae, P.K., Translation [rOtTI KunststojJe German Plastics, 76, 9 (1986). Bryk, M.T., Degradation of Filled Polymers, High Temperature, and Thermal-Oxidative Processes, Ellis Honrood, Chichester, (1991). Rabello, M.S. and White, l.R., Po(vmer Composites, 17, 5 (1996). Davidson, D. and Stewart, P., Society of Plastics Engineers ANTEC Tech. Papers, 31, 977 (1985). Hu, X., Xu, H., and Zhang, T., Polymer Degradation Stability, 43,225 (1994). Nikolova, M. and Mateev, M. Polymer Degradation Stability, 43, 977 (1985). Narisawa,1. and Kuriyama, T., AngelV Macromol. Chemistl)!, 216,87 (1994). Casu, A. and Gardette J., Polymel; 36, 4005 (1995). Goldman, A.Ya, Sorokin, A., Eisenhour, D., Barajas, A., Montes, J.A., and Beall, G., Intern't Conference on Polymer Characterization, Univ. of North Texas, Jan. (1997).

Effect of Aging

12. 13. 14. 15. 16. 17. 18. 19.

Murthy, N.S. and Minor, H., PoZrmeJ; 31,996 (1990). Cerius User Guide. Molecular Simulations Inc., San Diego, (1997). Peak Solve-Peak Fitting for Windo\vs, User's Guide. Galactic Industries Corp. (1991-95). Garbauskas, M.R., LeGrand, D.G., and Goehner, R.P., Advances in X-Ray Analysis, 36, 373 (1993). Wypych, G., Handbook of l\1aterial \Veathering, 2nd edition. Chem Tec Publishing, Toronto (1995). Gooden, R., Davis, D.D., Helln1an, M.Y., Lovinger, A.J., and Winslow, F.H., Macromolecules, 21, 1212 (1988). Minkova, L. and Nikolova, M., Polymer Degradation Stability, 25, 49 (1989). Schurz, J., Zipper, P., and Lenz, S., 1. Macromol Sci., Pure Appl. Chem., A30, 603 (1993).

209

The Influence of Degraded, Recycled PP on Incompatible Blends

Chiudia M. C. Bonelli, Agnes F. Martins and Eloisa B. Mano Instituto de Macronl0leculas Professora Eloisa Mano, Federal University ofRio de Janeiro, 21945-970 Rio de Janeiro, RJ, Brazil

Charles L. Beatty Departnlent ofMaterials Science and Engineering, Universit)) ofFlorida, Gainesville, FL 32611, USA

INTRODUCTION The recycling of post-consumer polyolefins has attracted much interest since these resins low density polyethylene (LDPE), high density polyethylene (HDPE) and polypropylene (PP) - are some of the most common polymers in the domestic plastic waste stream. 1,2 About half the weight of the total polYlner produced in the world is composed of polyolefins. Those are the cheapest plastics and are largely used for packaging. Their impact on the environment is considerable due to the low density and the hollow shape of the one-way packaging items, like bottles, containers, bags, etc, which contribute for them to en1erge either in waters or landfills. The large volulne they occupy makes them more conspicuous as nature pollutants than other waste products of equivalent weight. Recycled plastic can be obtained by two different approaches: a two-step process, involving plastics fractionation and processing of the separated plastics fractions; and a single-step process, using directly the mixture of plastic residues. Reprocessing mixed polyolefin waste can lead to products with lowered mechanical properties, since these polymer mixtures are usually incompatible. 3 Due to its favorable characteristics of price, density and versatility, PP is gradually replacing some materials. Its incompatibility with LDPE and HDPE causes loss of the mechanical properties. For example, HDPE or LDPE items may contain PP caps as contaminants, which are difficult to completely separate from the other polyolefins due to their similar densities and physical properties. 4,5

212

Weathering of Plastics

The use of compatibilizing agents may overcome this problem to a certain extent. Compatibilization ofthe Inultiple composition mixtures offers the possibility of reversing the deterioration of properties as less sorting occurs. 6 It can be commonly attained in principle through melt processing techniques, using in situ fonned copolymers, or adding copolymers, or low molecular weight con1patibilizing compounds. The CopolYlners have Segn1entS capable of specific interactions and/or chemical reactions with the blend components. 7 In this paper, we investigated the compatibilizing action of molecularly modified, recycled PP on the mechanical properties of SO/50 PP/I-IDPE blend, as suggested by preliminary experiments carried out in this laboratory.

EXPERIMENTAL MATERIALS The raw materials used in this work were: • PP, supplied by PPH Companhia Industrial de Polipropileno, Rio Grande do SuI, Brazil; type H503; specific gravity, 0.90 g/cm 3; MFI, 2.9 gilD min; • HDPE, supplied by Polialden Petroquimica, Bahia, Brazil; type BT 003; specific gravity, 0.95 g/cm3; MFI, 0.3 gilD min; • Post-consumer rigid plastic waste, supplied by the Municipal Company of Urban Solid Waste - COMLURB, Rio de Janeiro, Brazil- 50 kg.

METHODS The post-consumer raw n1aterial for the preparation of recycled PP (PPrl) was submitted to cutting, washing with water and drying in industrial equipment. A representative sample of the resulting flakes was taken for use in this work. Fraglnents ofPP were separated from other polymers by floating successively in water and hydroalcoholic solution (sp. gr., 0.91 g/cm 3) in tanks of200 liters and dried at room temperature (30°C).2,3 Binary and ternary blends were prepared in a Brabender single-screw extruder, model GN F 106/2, with L/D=25 and screw diameter 19 mm; screw rotation speed, 100 rpm at 190,200,210 and 215°C. The extrudates were cooled at 25-30°C and reduced to particles under 2.7 mm length. For the ternary blends, recycled PP was incorporated on a basic 50/50 PP/HDPE mixture. PPrl was ground and extracted by methyl-ethyl-ketone for 60 hours. The extracted recycled material (PPr2) was also incorporated to the binary mixture. Molecular weights (Mw ) were determined by GPC in a Waters 510 equipment, with differential refractometlY 410 detector, using trichlorobenzene for polymers and chloroform for extract as effluents. IR spectra were taken in FTIR Perkin-Elmer 1720 spectrometer. Solid-state NMR spectra were performed in a CP-MAS, Varian VXR 300 equipment, frequency 75.4 MHz, pulse 90°. DSC data were obtained in a Perkin-Elmer model DSC-7

Recycled PP

213

Table 1. Physical, thermal, and mechanical analyses of virgin and recycled polyolefins Test M\\ M\)M n Tm,oC Tc,oC Tonseb °c Tensile strength at Stress, MPa yield Elongation, % Tensile strength at Stress, MPa rupture Elongation, 0/0

HDPE

PPrl

PPr2

172,500 5.0 166 117 440 36 13 35 640

277,000 8.9 139 121 470 29 11 21 630

38,500 3.5 163 128 443 28 14 27 523

45,500 2.2 163 128 436 29 9 29 601

equipment, using IOo/2°C/min heating/cooling rate. The melting temperature (Tm) and crystallization temperature (Tc ) analyses were run from 30 to 200°C. TGA analyses were carried out in a Perkin-Elmer 7 Series Thermal Analysis System, using 1DOC /min heating rate under nitrogen, from 100 to 550°C. Melt flow index (MFI) tests were performed according to ASTM D1238, procedure A, conditions E and L, in a Emic equipment IFT-315. Specific gravity measurements were taken according to ASTM D792. Tensile tests were carried out according to ASTM D1708 in an Instron tensile tester, model 4204, 1 kN cell, cross-head speed of 1 cm/min, gauge length of 2.225 cm. Samples were cut from 0.1 x 15.0 x 15.0 cm plates, molded in a Carver press at 200°C and 22.2 kN, for 5 minutes.

RESULTS AND DISCUSSION The total loss in the industrial grinding, washing and drying was about 10% in weight. PP fraction, obtained by sink-float procedure, represented 10% of the total fragments of polyolefin residues. The extraction of polar contaminants from recycled PP (PPr1), cOIning from unremoved food residues (oils and fats), resulted in only 1.5 % waxy extract ( Mw = 725, Mw/M n =1.2), remaining apparently unaffected powder residue (PPr2). The extract shows monodispersity, which could be expected for non-polymeric nlaterial. Table 1 presents the results of the physical, thermal and mechanical tests perfolmed on virgin and recycled polyolefins. The nl01ecular weight ofPPr1 was lower than the virgin PP, since the recycled n1aterial may have been exposed to environmental and thermal degradation involving mainly n1acromolecular chain cleavage, probably with oxidation to some extent. Consequently, there was an increase in the melt flow index (14.8 g/l 0 min) and a decrease in the tensile strength at

214

Weathering of Plastics

Ii I'

I I .

I .

.. ' t'

,",.

14'

III

~"" ..,,_ , "._,v,_.'. ,., 141

" ,,' .f ,', "" ".u." ..

,...•.. , •

,,,

.,.

. ~

-0'

,., "", • .,

.. ·.··"1·· .. '··· " t ' ' ' ' , •. l'

••

1)1\11

j\J '. V\ ,__,

"·····'.~··

_._

" ..•:••' .... "; ..-, .. ':;.

_" ." -tt

Figure 2. NMR spectrum ofPPr2.

Figure 1. NMR spectrum of PPr1. lO'

- r - - - - - - - - - - - - -.. . .

-,.'

. -------------.....,

r .,:. 1 M"

Of-

n·

... .

o. _

.. Figure 3. IR spectrum ofPPrl.

, ~,M

ute

Jot. u"

ltQ•

Figure 4. IR spectmm of PPr2.

yield (28 MPa) ofPPrl, as compared to virgin PP (2.9 gilD min and 36 MPa, respectively). PPr2 did not exhibit significant differences in the molecular weight and in the tensile strength, as compared to PPr 1. T m and Tc confinned the composition of PPr 1. The temperatures of degradation (Tonser) ofPPrl and PPr2 were the same as for virgin PP. HDPE data showed that the higher degree of crystallinity was parallel associated to the higher Tonser.

Recycled PP

215

Table 2. Thermal and mechanical analyses of polyolefin blends PP/HDPE/PPr2

PP/HDPE/PPrl

Test

50/50/0

50/50/1

Tm,oC

165; 136

165; 135

164; 135

165; 135

163; 134

164; 135

Tc,oC

122; 128

122; 125

121; 125

122; 124

Tonseh °c

50/50/2

50/50/5

50/50/2

50/50/5

360

452

444

456

454

450

Tensile strength at Stress, MPa yield Elongation, 0/0

Tensile strength at Stress, MPa rupture Elongation, %

C~n I, leiA " II ~~hll ~B8 5o-p1# }l"'~t, •• A~9

!ted DI:~ 'll I~, 4l*~ 1997

Figures 1 and 2 show NMR spectra of PPrl and PPr2, while ' t c50·S(1..2X ! I ISO.O Figures 3 and 4 present IR specI ~ ....." .... " 1 i .. • 3~/~P,r(~-50-1 I .ISf't: Xl I I tra of PPrl and PPr2. These 0 ~~~~.(3;;~U-3. spectra showed the characteris~.J._._ 1. ,~.m~~U~.,·w2) "Q 100.0 .;~ ~ hdpaIPF"'ht tic peaks of PP, as expected. On -5' - ~o~.~> '~ the other hand, NMR spectra of \ PPrl and PPr2 exhibited an adi\ ditional peak at 33 ppm, \ 2:5. 0 associated with unsaturated ~:~'f C.D ethylenic, probably vinyl termiI I nal group and/or vinyline units, 1 200.0 100.0 300.0 '00.0 at the middle of the chain. IR N, ~~~: ~D:S ~ ~l=l: spectra of the recycled materials did not show any carbonyl abFigure 5. TGA analyses of polyolefin blends. sorption, which evidences that the PP degradation was not oxidative, with chain cleavage and unsaturation which did not change the apolar character of the PP molecule and kept its affinity to other polyolefins. 8 Table 2 shows the results of the thermal and tnechanical analyses performed on polyolefin blends. Concerning thermal data from the virgin polymers and their binary/ternary blends, DSC measurements indicate that the differences in Tc were larger than in T m. HDPE, which has higher degree of crystallinity than PP, was more affected by the presence of PP. TGA performed on polyolefin blends showed results presented on Figure 5. It can be inferred that there was an increase of 100°C in Tonset with addition of 1 % of PPr1 and/or PPr2 to 50/50 PP/HDPE blends, indicating some compatibilizing action of the recycled materials on the blends. "'9

P"f~lJ!('O:SO)

*11"'/~f50/~

..,,,.,t.:

(y\ •

.:>

J ___

i:'p/~cP'l*'~

~e.

14!~.

r='=

1--.-

\\',

'~'i i 1\'"

~o.o

216

Weathering of Plastics

•

l"J'lfitwE~~

•

lP'ti(W£1l!"tl

:·U·~~I

P,til»t1'f"tl

~!4'2

"'H09"E"P'rI~J4"

PI' HOX PM ~o..~..oJ f>PHWf.:''Pfl!~ ~~,~~

(;

·tu

Figure 6. Tensile strength tests of polyolefin blends.

The mechanical prope11ies of the polyolefin blends are shown on Figure 6. It is possible to see the incompatibility of 50/50 PP/HDPE blend, causing the premature break of the test samples before reaching the yield point. The increase of tensile strength and elongation at rupture of ternary blends as conlpared to binary blend indicated some compatibilizing action ofPPrl on the system conlponents. The ternary blends broke before the yield point. Small quantities ofPPrl were enough to produce better results on the mechanical properties of the blends under investigation. The addition ofPPr2 on binary blends was more effective, as far as compatibilizing action was concerned. The ternary blends reached and surpassed the yield point before breaking.

CONCLUSIONS The experimental results suggest that the degradation which occurred in PP molded, post-consumer artifacts after exposition to natural, uncontrolled outdoor conditions provided spontaneous, non-oxidative chemical modifications on PP molecules which brought a certain degree of compatibilization action towards polyolefin residues.

ACKNOWLEDGMENTS The Authors thank the Army Research Institute, Rio de Janeiro, RJ, Brazil, for the DSC analyses.

REFERENCES 1 2 3 4 5 6 7 8

c-s. Ha, H-D. Park, Y. Kim, et aI., Polymers for Advanced Technologies,

7, 483-492 (1996). C.M.C. Bonelli, "Recuperacao Secundaria de Plasticos Provenientes de Residuos So 1idos Urbanos do Rio de Janeiro", M. Sc. Thesis, Instituto de Macromoleculas, Federal University of Rio de Janeiro, Brazil (1993). E.B. Mano, C.M.C. Bonelli, M.A. Guadagnini, and S.J. Moyses-Luiz, Polimeros: Ciencia e Tecllologia, 4 (3) 19-24 (1994). R.S. Stein, "Miscibility in Polymer Recycling", in Emerging Technologies in Plastics Recycling, 513, 3948, ACS Symposiuln Series, Washington (1992) C.L. Beatty, Proceedings of SPE Annual Technical Conference, 3032-3033 (1994). K.C. Johnson, Proceedings of SPE Annual Technical Conference, 3732-3737 (1995). A.L. Bisio and M. Xanthos, "How to Manage Plastics Waste - Technology and Market Opportunities", Hanser/Gardner, New York (1994). Y. Long, B.E. Tiganis and R.A. Shanks, J. Appl. Polym. Sci., 58, 527-535 (1995).

Interactions of Hindered Amine Stabilizers in Acidic and Alkaline Environments

K. Keck-Antoine and D. Scharf Specialty Che111icals Group, BU Additives, Hoechst Celanese Co/po Charlotte, NC 28217, USA H. Koch R&D Departnlent, BU Additives, Hoechst AG; Augsburg, Gerl1zany

INTRODUCTION Hindered Amine Stabilizers (HAS) are very effective UV-stabilizers that outperfonn all other types ofUV-stabilizers mainly in polyolefins. In addition some HAS are known to offer outstanding long tem1 thermal stability. This high efficiency is based on a radical scavenging mechanism. However, cases have been reported where the performance of HAS was significantly lower than expected. In the majority of these cases, HAS stabilized polyolefin films had been in contact with reactive chemicals and subsequently failed prematurely. While the chemical reactivity of HAS is needed for their outstanding performance it can cause antagonistic interactions in the presence ofother reactive chemicals. These interactions can significantly decrease the UV-performance of HAS. In addition, interactions between HAS and reactive chemicals can also influence the long tem1 thermal stability, processing stability and discoloration effects of polyolefins.

ALKALINITY OF HAS Hindered Amine Stabilizers (HAS) are basically radical scavengers which require a certain level of chemical activity. As a result of their amine chemistry, they can be expected to be more or less alkaline. Very often the pKa-value is used to characterize the alkalinity of HAS (Table 1).

218

Weathering of Plastics

Table 1. Alkalinity of HAS (HMW = high moOne possible reaction scheme of lecular weight, LMW low molecular weight) acid-HAS interactions describes the salt fonnation as a result of an acidpKa base reaction. 4 Such a salt fonnation HAS Type [1] [2] [3] would deactivate the functional HAS-l HMW 9.7 9.2 8.6 group of the HAS and consequently HAS-2 9.2 HMW. 9.1 limit its perfolmance. HMW HAS-3 6.5 6.5 5.5 During processing, storage and HMW HAS-4 9.6 use, HAS-stabilized polymers may HMW HAS-5 6.7 be exposed to more or less strong acLMW 9.0 HAS-6 9.3 ids or (more general) reactive LMW 9.2 HAS-7 chemicals which can migrate into the polymer. Further, acids or reactive Table 2. Influence of acid exposure on chemicals can fonn in the polymer HAS-stabilized LOPE films matrix or can already be present due to other additives or ingredients.

Acid

UV-Stabilization improvement factor

none

1.0

HAS-l >15

HAS-5 >15

HN0 2

1.1

6.8

10.3

H 2S03

0.6

1.4

2.1

san1ple: 300 micron blown film; stabilization: LDPE-l + 3000 ppm HAS; treatment: dipped each 100 h for 16 h in 0.1 n acid, washed with deionized water and dried at room temperature; criterion: exposure tin1e until ~CO=O.3; weathering: X 150 xenon-arc (standard conditions).

UV·STABILITY OF POLYOLEFIN FILMS A typical example are agriculture PE films used for crop enhancement. These films are in contact with reactive chemicals 4 and often show "un-

explained early degradation" under field conditions. In a model experiment, LDPE films containing HAS with significantly different alkalinity were brought into contact with two different acids (Table 2). As predicted from the acid-base reaction, the stronger acid, H2S0 3 (pKa == 1.92) had a more negative impact on the film perfonnance versus HN0 2 (pKa == 3.34). The film with the more alkaline HAS-l (pKa==8.6)3 was significantly more affected by either acid compared to the less alkaline HAS-5 (pKa==6.7).3 Without acid contact, both films revealed a comparable lifetime. To confinn salt formation as one possible mechanism the experimental set-up was repeated and, additionally, the accumulation of selected trace elements in the films was measured. 4 Trace elements were sulfur for H2 S0 3 and Metham Sodium respectively chlorine for Sumi(sc)lex (Tables 3 and 4). In all three cases a correlation was found between the perfonnance of HAS and the accumulation of ce11ain trace elements in the film. The films with the less alkaline

219

Interactions of Hindered Amine Stabilizers

Table 3. Activity of HAS after contact with reactive sulfur containing chemicals

Chemical

H2S03 (0.1 mol/I)

Metham sodium (3% solution)

Stabilization 5000 ppm HAS-l 2500 ppm UVA-l 5000 ppm HAS-5 2500 ppm UVA-I 5000 ppm HAS-I 2500 ppnl UVA-l 5000 ppIn HAS-5 2500 ppm UVA-l

Retained relative elongation [%] after after 2000 h 74

Sulfur content [ppm] after 2000 h

Sulfur increase (linear regression)

Exposure tiole [h] until 1000 ppm sulfur in the film

750

y=0.24x+274.2

3050

444

y=O.16x+122.8

5462

1640

y=O.80x+31.1

1204

737

y=0.34x+45.8

2762

sample: 200 micron blown film; stabilization: LDPE-l + HAS + UVA-l; treatment: each 144 h contact for 24 h with chemical; dried at room temperature; criterion: retained relative elongation at break [%] sulfur content [ppm]; weathering: X 450 xenon-arc; standard conditions (no rain cycle)

Table 4. Activity of HAS after contact with reactive chlorine containing chemicals

Chemical

Sumisclex

Stabilization 1500 ppnl HAS-l 1500 ppm HAS-5

Retained relative elongation [%] after after 1000 h 14 68

Sulfur content [ppm] after 1000 h

Sulfur increase (linear regression)

Exposure time [h] until 1000 ppm sulfur in the film

335 279

y=0.32x+ 15.0 y=0.24x+39.1

3078 4004

sample: 200 micron blown film; stabilization: LDPE-l + HAS; treatment: each 125 h contact for 24 h \vith 0.05 Procymidon solution; dried at room temperature; criterion: retained relative elongation at break [0/0] sulfur content [ppm]; weathering: X 1200 xenon-arc; standard conditions (no rain cycle)

HAS-5 showed a longer lifetime and accumulated less trace elements. This means that the exposure time to reach a threshold trace element level was significantly longer. The trace element accumulation showed linear behavior.

220

Weathering of Plastics

As reported earlier5 there was further evidence of the potential deactivation of HAS due to salt formation. The in-situ formation of [HAS-1 ]sulfite showed an IR-absorption at 2480 cm- I •3 Additionally, films containing in situ created [HAS-1 ]sulfite showed no peaks at 1565 and 1530 cm- I compared to films containing "standard" HAS- 1. Shachar et at. found similar phenomena. 6 Although acid-HAS (base) reactions are a significant part ofpotential antagonistic interactions, it seems that other mechanism may occur as well with complex chemicals.

"LONG TERM THERMAL STABILITY" OF HOPE GEOMEMBRANES Most of the work related to acid-HAS in-

Table 5. Onset of Oxidation (OIT) of teractions has been focused on thin-walled HOPE plaques after contact with reactive

chemicals (OIT = Oxidation Induction applications (films) and the performance

of HAS as UV-stabilizer. Some of the interactions appear to require the presence ofUV-light. 3 OIT, min However, they may also occur in none HAS-l HAS-5 thick-walled polyolefin applications, without acid 36 63 67 without excessive UV-light and may afwith acid 29 56 63 fect the long tem1 them1al performance. sample: 1 mm injection molded plaques; stabilization: HDPE-2 + The lifetime of HDPE plaques was meaAO (proprietary) + carbon black + 900 ppm HAS; treatment: sured by DSC (OIT value). When no acid dipped once into 0.25 mol/l H 2S03 for 48 h; criterion: OIT at 210°C [min] is present HAS serve as excellent long term thermal stabilizers. However, the contact with H2 S03 decreases the OIT value. This decrease is more pronounced with the higher alkaline HAS-l than the less alkaline HAS-5 (Table 5). Time)

PROCESSING OF POST CONSUMER POLYOLEFIN RESINS An interesting effect was observed when processing post consumer polypropylene battery case resin into multifilament yam. Objective of the study ofG. Coy7,S was to efficiently produce quality fiber from recycled raw materials of different sources. The maximum take-up speed was found to correlate closely with observations made during actual processing and can be used as criterion for spinnability. The addition of an antioxidant package increased the spinnability of the compound. Further addition of UV-stabilizers of the HAS type (in particular HAS-I) either maintained the spinnability level or even increased it.

Interactions of Hindered Amine Stabilizers

221

Table 6. Spinning stability of post-consumer polypropylene This result (Taresin. 7,a Maximum take-up speed [m/min] as function of the ble 6) contradicts additional stabilizer package experitnents with a different set-up as described elsewhere. 9 The addition of HAS-l decreased the spinnability beCar battery resin 299 340 273 583 low the level of the waste control sample stabilization: PP-l (50%) + post-consumer PP (50%); criterion: "spinnability", maximunl without any additake-up speed [m/min]; equipment: Research Spin Unit (RSU) [customer results] tional stabilizer. The battery case resin contained (despite cleaning) about 200 ppm of sulfur; indicating acid contamination. Consequently, an interaction between the acid and the highly alkaline HAS-l was expected to take place. This was confirmed by repeating the experiment with the low alkaline HAS-5, which resulted in a strong improvement of spinnability. The poor improvement in UV-stability of the formulation containing the highly alkaline HAS-l provided further evidence of acid-HAS interactions and a corresponding deactivation of the HAS-structure. "Spinnability" Take-up speed as function of stabilizer package 0.04% AO-I, 0.16% P-l, 0.04% AO-I 0.20/0 HAS none O.16%P-l HAS-l HAS-5 Bottle resin waste 207 251 not tested 211 rm/n1inl

DISCOLORATION OF ADDITIVE CONCENTRATES IN ALKALINE ENVIRONMENTS Table 7. Discoloration of additive con- Certain additive concentrates, like those typicentrates. Color deviation after 375 cally used for LDPE agriculture film, may days of storage at room temperature contain polyn1eric HALS, UV absorbers and HAS conlponent in Concentrate

Total color change ~E after 375 days

HAS-l

HAS-5

42.9

4.1

formulation: 81.0% LDPE-l + 12.0% HAS + 6.0% UVA-l + 1.0% AO-2

phenolic antioxidants. Concentrates containing UV absorbers are yellow from the beginning due to the absorption of the benzophenone (or benzotriazole) structure. In contrast to concentrates containing HAS-5, which show little to no discoloration, concentrates based on the highly alkaline HAS-l discolor strongly (Table 7). This continuous discoloration with HAS-l (and

222

Weathering of Plastics

oo~--------------,

HAS*11

• •• • ··HAS-S

Q.J-_ _..........AIIIil:....-.,...a=--IIIIIIC:l:;.;----i 2CO

2&J

4CX)

wave length [nm] Figure 8. Light transmittance ofLDPE films. Sample: approx. 80 micron compression molded film; formulation: see Table 7. FilIns were manufactured from discolored additive concentrates (Table 7).

mostly all highly alkaline HAS) is less pronounced in the absence of either the benzophenone structure (UVA-I) or a hindered phenolic structure (AO-2). The color shift does not occur in the absence of both, the UV absorber and the hindered phenolic structure. 3 In this case it appears, that the HAS structure is not directly involved in chemical interactions. Rather it seems that more alkaline HAS structures create a sufficiently alkaline environment which favors the oxidation of phenolic structures thus creating

co ore qUInone structures. Films compression molded from the concentrates (no letdown) did not indicate a lack of active benzophenone. This is probably due to the very high overall amount ofbenzophenone in the film. However, analysis of the active benzophenone content gave indications of a loss in the range of 10% (Figure I). On the other hand it was previously reported 10 that a difference in light transmittance in the range of 350 - 400 run has a significant impact on the quantity of light available for the crop and consequently the corresponding crop yield.

CONCLUSIONS The high alkalinity of son1e HAS structures is responsible for interactions that can reduce the performance of HAS. Evidence was found that this is not only true for UV-stability, but for long teon thermal stability and processing stability as well. Additionally, certain discoloration phenomena appear to be caused by the alkalinity of HAS. Consequently, the selection of the HAS structure should take into consideration the perfonnance under non-ideal, e.g. acid exposure conditions and should not only focus on the UV aspect. In particular in applications where interactions are quite obvious, the selection must focus on the perfonnance ofHAS and the potential risk of premature failure due to HAS deactivation.

REFERENCES 2 3

Horsey D., Leggio A., Reinicker R.; Hindered amine light stabilizers (HAS)/pigment interactions -HAS structural effects on color strength; published by SPE (Effects in Plastics); 1993. Gray R.; A novel non-reactive HALS boosts polyolefin stability; Plastics Engineering; June 1991. Hoechst AG, inteolal data.

Interactions of Hindered Amine Stabilizers 4 5 6 7 8 9 10

223

Keck-Antoine K.; Stabilization ofa*griculture filnls by polynleric HALS with particular emphasis on possible interactions \vith agro - chelnicals; ANTEC '95; Boston, MA. Carlsson D., Can Z., Wiles D.; Polypropylene photostabilization by hindered amines in the presence of acidic species; Journal ofApplied Po(rmer Science, Vol. 33,875-884, 1987. Shachar R., SteIman R., Shai E., Efrat B., Ashkenazi Y., Asenheirn D.; HALS stabilized LDPE agrifilms under the influence of elenlental sulfur, 1996. Coy G.; Processing post-consumer poly-propylene resin; Sunlmer Intern, Virginia Polytechnic Institute and State University; 1995. Coy G.; The properties, morphology and stability ofmultifilament polypropylene yam containing post-consumer recycled resin; Virginia Polytechnic Institute and State University, 1994. Keck-Antoine K.; Interactions of hindered amine stabilizers - During processing and manufacturing; Additive '97; New Orleans; February 1997. Lagier 1., Rooze K., Moens F.; COlnparative agronomical experiment on greenhouse filnls stabilized with HALS and nickel quenchers; Plasticultllre, #96; 1992.

CODE

TRADENAMF,

SUPPIJER

HAS~l

1t) Chimassorb ~.44

elBA Additives

Chimassorb 119

elBA Additives

HAS-J

® Tinuvin 622

elBA Additives

HAs....

® Cyasorb 3346

STRl;C'fURE

Weathering of Plastics

224

eoof.

TR.-\DENAME

SUPPLIER

HAS-5

Hocchst AG

Ho!t3\'in NjO

STRUCTURE

CH. CH,

CH.-tCH"'l I CH.

H-N

C-N-R

CH, CH, ~

HAS-6

Tin",in 770

elBA Addilivcs

H~O-C-ICH,lrc -O~NH

>.:....J

II 0

\....,('

HoC-tCH ,» HAS-7

Hoslavin N20

CH, CH,

Hooch$! AG

I o-_ .... -CH.

H··- N

C····N-H CH, CHl I!

° UVA·l

Chillla,som 81

ellA Adduivcs

Hostavin ARO 8

H~hSl

(( ~

A().I

li)

A().2

Irgano, 1076

1'·1

1')

Irganox 1010

IrCi'fos

J6~

)\. HostallO.\. PAR24

'----

~,&OH

OCaHu

elDA Addih"<S

elBA Addili,·cs

ClIlA Addlll\'CS

Hocch~l AG

V -

I' CH,CH,COC"H..

rL4 ~"\- o_l, _f)· _.

I _.---.J

Appendix I. Agrochemicals used in the study

Trade name

Type

Application

KMetham Sodium

sodium methyldithiocarbamate

soil fung,icide, nematicide, herbicide

KSumisclex (Sumilex)

N·(3,S·dichlorophenyl)·I,2-dimethylcyclopropane·I,2dicarboximide

fungicide

Interactions of Pesticides and Stabilizers in PE Films for Agricultural Use

Edina Epacher and Bela Pukanszky Technical University ofBudapest, Depart111ent ofPlastics and Rubber Technology, Institute ofChe111istlY, Che111ical Research Centel; Hungarian Acadelny ofSciences

INTRODUCTION Traditionally Hungary is an agricultural country. In recent years the use ofPE films for greenhouses became widespread, the production of such films increased significantly. Proper stabilization of films used for such a purpose is an important financial issue, a more efficient stabilizer package extends the lifetime of the film, but increases its price. On the other hand, the cost of installing and dismantling the houses, as well as that of the disposal of waste films decrease if the film lasts several seasons. However, the stabilization of agricultural films is a serious technical challenge. During their use, the films are exposed to the effect of oxygen, tTIoisture, summer heat, and UV radiation, among which the last has the strongest influence on lifetime. As a consequence, the most important, or even the sole aspect of stabilization in this field is the development ofan appropriate light stabilizer package. PE grades used for the production of agricultural films practically always contain a hindered amine light stabilizer (HALS) and often also an UV absorber. Properly stabilized films survive two, sometimes three agricultural seasons. However, antagonistic interaction of light stabilizers and phenolic antioxidants was observed sometimes, which may decrease the efficiency of the stabilizer package. 1,2 Further interactions are expected in the presence ofpesticides which are used for the protection of the crop grown in the greenhouse. Usually pesticides have complicated formulations, they contain a number of compounds beside the active component. It is a well known fact that films are destroyed prematurely when certain pesticides are used indicating an antagonistic interaction of the formulation and the stabilizer package. The pesticide must react with the stabilizer decreasing its effect or conlpletely destroying it. The practical importance of the problem is obvious, thus the goal of our study was to identify the pesticide

226

Weathering of Plastics

formulations or active components which decrease the lifetime ofPE films, on the one hand, and to grade the stabilizer packages tested, on the other. Furthermore, an attempt was made to explain the mechanism of interaction in the presence of harmful formulations.

EXPERIMENTAL

MATERIALS The same PE grade (Tipolen FA 2210, TVK, Hungary) was used throughout the experiments. The performance of three stabilizer packages was compared in a film with an anticipated lifetinle of 1 year. Stabilizer package S, the standard system ofthe producer of the film, contained a combination ofTinuvin 622 and Chimassorb 81 UV. The experimental package A consisted ofHostavin N30 and Hostavin ARO 8, while package B ofTinuvin 622 and Chimassorb 944. Tinuvin 622, Hostavin N30 and Chimassorb 944 are HALS compounds and Chimassorb 81, which corresponds to Hostavin ARO, is an UV absorber. An attempt was made to use the widest possible range of pesticides. The formulations used in the largest quantity in Hungary were all included into the study and practically the complete range ofactive cOlnponents were represented among the investigated products including organic phosphorous and sulphurous compounds, halogenides, organometallic compounds, etc.

MEASUREMENTS The selected 24 pesticides were diluted with water to a concentration recommended by the supplier for the user. The films were soaked in these solutions for 1 hour, 1 day and 1 week. The first corresponds to a weak, the second to a moderate, while the third to a strong exposure to the effect of the pesticide. Weathering experiments were carried out under dry conditions for 300 and 600 hours with films treated for 1 day. Tensile properties were measured on small dumbbell specimens cut parallel to the extrusion direction. The measurements were carried out on a Zwick 1445 machine with 100 mm/min cross head speed. Thermooxidative stability ofthe films was characterized by the initial temperature of degradation (Td) measured in nonisothermal degradation experiments at 1O°C/min heating rate on a Perkin Elmer DSC-2 apparatus. FTIR spectra were recorded on a Mattson Galaxy 3020 apparatus in the range of 4000 and 400 cm- 1 wavenumbers, while UV spectra on a Hewlett Packard HPUV 8452 equipment between 200 and 800 run wavelengths.

RESULTS All c01l1binatiolls of24 pesticides, 3 stabilizer packages, and 3 soaking times represent a large number of experiments and measurements, thus some screening tests were carried out in a first step. In latter stages the effect of those pesticides was studied only, which considerably influenced the properties of the films. Our attention is focused mainly on sulphur containing

Interactions of Pesticides and Stabilizers

............. .-- ---,--- --I D 100'..>1

.1<'-,,.,. I Dl1(,'E"'C"~:

227 formulations in this paper, because these showed the most contradictory behavior and deteriorated the properties of the films in the largest extent.

OXIDATIVE STABILITY The measurement of the initiation of oxidative degradation in non-isothermal experiments is a convenient and quick method for the detection of chemical interactions between the pesticides and the stabilizer package of the film. Figure 1 presents the change in the initial temperature of degradation of the film containing the standard Figure I. Effect of pesticides and time of soaking on the oxidative stabilizer package (S) with varying soaking stability ofa PE film stabilized with package S. time for selected pesticides. Stability decreases already when the film is soaked in distilled water. The effect can be caused by the dissolution of one of the components of the stabilizer package or by the hydrolysis of a functional group. Ifwe consider the fact that the films are exposed to rain and watering of the crop, such an effect is highly undesirable. Soaking in water did not decrease stability when the other two stabilizer packages were used (A, B). The contact with pesticides results in considerable changes in the oxidative stability in both direction. Copper, carbamate and thiocarbamate compounds decrease stability, while pesticides containing elemental sulphur as an active component improve it. This might be surprising at first, but we must bear in mind that sulphur and certain sulphurous compounds decompose hydroperoxides and act as stabilizers. 3 Pesticides influenced the stability of the films containing the other two stabilizer packages in a similar way, with the only difference that the magnitude of stability changes were smaller in both cases. Although the observed increase in stability is advantageous, it does not necessary mean that sulphur containing pesticides improve the lifetime of the films. Under service conditions the main factors influencing stability are UV radiation, oxygen, moisture, etc, when both initiation and degradation reactions, as well as their rates are different. As a consequence, caution must be applied when conclusions are drawn from these results.

SPECTROSCOPY Spectroscopy can detect chemical interactions between the pesticides and the stabilizers, but also physical changes, i.e. the diffusion ofa component into the film or the opposite effect, its

Weathering of Plastics

228

Absorbance 2. 15S'

2.2519

1.7se7

1.2556

O.75i53

200

ioo

300

Wavelength (nm) Figure 2. Changes in UV spectrum of a PE film containing package B upon soaking for various tin1es in a pesticide of moderate interaction (Chinetrin 25WC, CI&N).

intennediate interaction led to the change of the intensity of absorption bands without changing their wavelength (Figure 2), while strong interaction drastically modified the complete spectrum (Figure 3). Such a drastic change could result from the dissolution of an active component, diffusion of a con1pound

Absorbance 1 day

~lhOur Untreated t .. H$l

O.91l~3

0.2'1713

dissolution. PE does not absorb UV light, thus any observed absorption must be related to the amount and chemical structure of the stabilizer. Under the effect of irradiation chromophores (unsaturations, carbonyls) may fonn in the polymer which can also absorb light. Based on the changes in the UV spectrum of the filn1s pesticides could be divided into three groups. Inert fOlIDulations caused only slight changes of the spectlum at most,

l~~~~~~~~~~~~~~~~~= ~~ilierummmaybeilieco~

200

500

Wavelength (om) Figure 3. Effect of soaking time on the UV spectrum ofa PE film (package B) in contact with a strongly interacting pesticide (Acellic 50EC, S&P).

sequence of chemical reactions. Reaction products might have their own adsorption in the spectrum. According to this classification the three stabilizer

packages showed considerable differences in activity. We detected the highest number of strong interactions, which was 9 with package S, while the lowest with package A, i.e. three strong interactions. Mostly sulphur containing compounds and an organic halogenide (Neoron 500 EC, Br active component) entered into strong interaction with the stabilizers. Strong interaction forecasts sensitivity to UV irradiation, as well as more drastic changes in chemical structure and mechanical properties. This assumption was corroborated by the analysis of the spectra recorded after irradiation in the Xenotest, the spectra

229

Interactions of Pesticides and Stabilizers

Table 1. Carbonyl development in PE films stabilized by packages A and B as a function of irradiation time Pesticide Trade name \vater Acellic 50EC Chinetrin 25WC

Active component

S&P Cl&N

Relative carbonyl content after Xenotest time, h Package A Package B 300 600 300 600 0.52 0.66 0.20 0.31 0.46 0.39 0.39 0.55 0.29 0.23 0.46 0.49 8.23 0.44 0.43 -

Neoron 500EC

Szulfur 900WF

Se Se

1.04

0.70

2.14

9.47

Thiovit

0.90

2.83

Topas

0.43

0.49

0.28

0.33

Neviken

3.75

8.99

3.76

10.88

did not change in the presence of inert pesticides, while a considerable decrease of UV absorber concentration was detected under the effect of pesticides with strong interaction. Because of the relatively high concentration of the stabilizers, their presence could be detected also by FTIR spectroscopy, in spite of the lower sensitivity of this technique. The same changes could be followed as with UV spectroscopy, but some additional observations could be made as well. Diffusion of the active component into the filn1 could be followed in some cases, e.g. a vibration characteristic for ester groups appeared in the film and its intensity continuously increased with time (Chinetrin 25WC, CI&N; Neoron 500 EC, Br; Actellic 50EC, S&P). As an effect of irradiation, strong oxidation of the filn1 could be observed in the presence of some pesticides. The effect is demonstrated well by Figure 4 where the increase in the intensity of carbonyl vibration is shown in the presence ofNeviken containing elemental sulphur (Se) in the polysulphide form as active component. Not only the intensity of the absorption, but also its wavenumber changes as a function of irradiation time. Oxidation was observed in the presence of several other pesticides as well, sulphur containing compounds were the most active in this respect, too. The intensity of carbonyl vibration detected in films stabilized by packages A and B is presented in Table 1 for selected pesticide formulations. The wavenulTlber of the characteristic carbonyl vibration changed between 1718 and 1713 cm- 1 and depended on the type of the functional group formed, i.e. ester, carbonyl, aldehyde or ketone. The table clearly shows that some formulations are completely neutral. Very high carbonyl intensities are recorded in other cases (Neviken). Package A proved to be the most stable, high degree of oxidation could be detected only in the presence of one pesticide. For some reason sulphur in the polysulphide fonn strongly catalyzes the degradation of PEe

Weathering of Plastics

230

Table 2. Effect of pesticides and irradiation on the ultimate elongation of PE films Elongation (%) after treatment for 1 day and Xenotest time (h)

soaking

327.5

Acellic

351.8 314.4 324.6 339.6 356.9 311.6 337.1

Chinetrin Neoron Szulfur Thiovit Topas Neviken

Package B

Package A

Pesticide

300 341.3 394.4 318.3 356.6 309.3

337.3 278.3

600 340.4

soaking

222.3 333.0 52.9 230.5

292.0 282.7 317.3 323.8

300 303.8 247.0 331.4 215.4 181.6

301.6 349.8 50.3

300.1 316.0 287.7

309.9 130.2

332.7

600 307.4 199.2 281.9 32.5 0 139.6 285.0 23.7

MECHANICAL PROPERTIES Tensile strength and elongation-at-break of the films were determined on films before and after irradiation. These propel1ies are inlportant for practice and especially elongation reflects changes in the molecular structure of the polymer very sensitively. Elongation-at-break values of films exposed to the effect of the pesticides listed in Table 1 are compiled in Table 2. The changes in tensile strength completely correspond to those in elongation thus they are not listed here. The results of the table indicate that a simple soaking of the film in the pesticides does not influence its mechanical properties. Irradiation, on the other hand, leads to considerable deterioration of strength and elongation in the presence of some pesticides, while leaves these properties unchanged upon treatment with others. Neoron and sulphur containing fornlulations are the most hamlful again, i.e. the compounds interacting strongly with the stabilizers according to UV and FTIR spectroscopy. A close correlation can be observed between the oxidizing activity of the pesticide and its effect on mechanical properties (see Table 1). Retention ofproperties is much better if stabilizer package A is used; even the standard recipe is better than package B.

DISCUSSION The experimental results clearly show that pesticides with sulphur containing active components enter into antagonistic interaction with the stabilizers. One simple explanation could be that the interaction of the pesticide and the HALS compound prohibits the latter in exerting its effect. This reasoning, however, does not explain the fact that thennooxidative stability in-

Interactions of Pesticides and Stabilizers

231

creases in the presence of sulphur containing pesticides. An unfavorable interaction of the HALS compound should lead to the decrease of stability anyway, since these additives exert their effect by fonning stable nitroxyl radicals which act as radical scavengers. 3 ,4 However, stability decreases only under the effect of UV irradiation. A more plausible explanation assumes that the sulphur/HALS interaction promotes the fOffilation of 1800 1700 aoo Wavenumber (cm'1) alkoxy and peroxy radicals. Figure 4. Carbonyl development in a PE film (package B) soaked in a strongly HALS compounds may react interacting pesticide (Neviken, Se). with these oxygen containing radicals, but the stabilizer is consUlned rapidly because of their high concentration. This explanation is corroborated by several facts and observations. The increase of thennal stability indicates the activity of sulphur in radical reactions. Films stabilized with a system containing an UV absorber (packages S and A) perfonn better in irradiation experinlents, because this additive absorbs a part of the radiation thus inhibiting sulphur in the production of reactive alkoxy radicals. Upon irradiation and intensive carbonyl absorption appeared in the FTIR spectra when the film came into contact with sulphur containing pesticides. Moreover, a high concentration ofalkoxy radicals leads to chain scission and the filnls treated with sulphur containing pesticides could be dissolved completely, indeed. Contact with the organic bromine compound (Neoron), on the other hand, produced a film of high gel content, i.e. the mechanism of degradation is completely different in this case. Finally, the attention must be called to the fact that the three pesticides containing elenlentary sulphur as active component have rather different effect on stability, thus the properties (dispersion state, molecular weight,S modification) ofthe sulphur or sulphurous compound plays also an important role in its activity, polysulphide being the most deleterious.

CONCLUSIONS The extensive study of the effect of pesticides on the stability and lifetitne of agricultural fihns showed that the combined influence of some active compounds, stabilizers, UV irradia-

232

Weathering of Plastics

tion and oxygen leads to the rapid deterioration offilm properties. Although a large number of pesticides do not interact with the stabilizer package, some sulphur containing compounds and organic halogenides initiate the oxidation of PE and lead to the fast deterioration of its mechanical properties. The extent of this negative effect depends on the molecular weight, dispersion, allotrope modification, etc. of elementary sulphur. The disadvantageous effect of these compounds was tentatively explained by the generation of a large number of alkoxy radicals due to the interaction of sulphur and the HALS stabilizer. Clear distinction could be made in the efficiency ofthe three stabilizer packages used, their performance changed in the improving order of package B, Sand A. According to the results and in agreement with the above presented explanation, the introduction of an UV absorber into the package considerably improves the light stability and lifetime of the film.

REFERENCES 1. 2. 3. 4. 5.

Chirinos-Padron, A.J., Po(vm. Degl: Stab., 29,49 (1990). Allen, N.S., Edge, M., Rahlnan, A., Chen, W., Shah, M., Holdsworth, D., Polym. Degl: Stab., 44, 249 (1994). Allen, N.S., in Developments in Polymer Chemistry-2, ed. Allen, N.S., Appl. Sci. Publ., London,1981, p. 239. Yongcheng, Y., Pol)'lJ1. Degl: Stab., 37, 11 (1992). Bigger, S.W., Delatycki, 0., J. Polym. Sci., Polym. Chem., 27, 63 (1989).

The Influence of Co-Additive Interactions on Stabilizer Performance

Robert L. Gray and Robert E. Lee Great Lakes Che111ical Corporation

INTRODUCTION Polypropylene fiber has shown dramatic growth during the last couple of decades.! A key to this success has been continuous improvement in its UV stability as a result of significant advancements in stabilizer technology_ Before the advent of today's generation of stabilizers, polypropylene fiber was perceived very negatively, often as a result of poor stabilizer performance. 2 In the absence of stabilizers, polypropylene has very poor UV stability_ The practical consequences ofunchecked exposure to UV radiation are: discoloration, surface crazing (formation of surface nlicrocracks), enlbrittlelnent, and loss of mechanical properties (elongation, impact strength, and tensile strength). The effect ofUV exposure can be significantly inhibited through proper selection of UV stabilizers. Light stabilizers can be categorized into four general classifications: screening agents, UV absorbers (UVA), UV quenchers, and Hindered Amine Light Stabilizers (HALS). The most effective class of light stabilizer for lnost applications is the HALS. These materials have been shown to function as radical traps, thus interrupting the radical chain degradation mechanism. 1 The cyclic stabilization mechanism proposed for HALS involves multiple regeneration of the active nitroxyl stabilizer. The surprising performance of HALS at relatively low concentrations supports this non-sacrificial mechanism~ HALS 1 was the first commercially available HALS. This product was dimeric and had a molecular weight of 481 g/nlol. Initial work showed outstanding results in polypropylene. Unexpectedly, the high volatility ofHALS 1 made it susceptible to loss during post extrusion heat treatments such as tentering. 4 This is presull1ably related to the relative high volatility of HALS 1 which is a result of its low molecular weight and its propensity to quickly Inigrate to

234

Weathering of Plastics

the polypropylene surface. This problem was addressed with the introduction ofhigh molecular weight HALS such as HALS3. Recent moves into the residential carpet market required the n1aintenance of strict control over undesirable color development through processing, weathering, and storage (gas fade). In certain systems, HALS have been shown to contribute to color development through an interaction with phenolic antioxidants. This discoloration can be accelerated by exposure of the fiber to NO x type gases which can be generated by warehouse equipment. This color development can be minimized by removal of the phenolic antioxidant or use of a tertiary HALS4 or HALS2. 4 HALS4 was introduced, in part, to address the discoloration concerns that can occur with secondary HALS such as HALS3 while maintaining a high level of UV stabilization activity. Co-additive interactions must also be carefully considered when fOlmulating polypropylene fiber stabilization packages. Pigments have been shown to either enhance or degrade stabilizer performance. Similarly, stabilizers can have both positive and negative interactions with pigments. 5 UV stabilization of polypropylene fiber containing brominated flame retardants has been the focus of intense technical efforts with only limited success. Acid generated by the flame retardants deactivate HALS, thus severely reducing the HALS' effectiveness. Due to this interaction, flame retarded polypropylene fiber has been generally restricted to applications requiring little or no UV stability. While some progress has been made, this application continues to be a true technical challenge. The gains achieved in light stabilizer teclmology for use in polypropylene fiber has been impressive. This paper will introduce a new generation ofHALS which pushes the UV stability of polypropylene fiber to new heights. This novel technology provides dramatic in1provements over the previous state-of-the-art in diverse areas such as pigment interaction, flame retardant stabilization, and of course, general UV stabilization.

EXPERIMENTAL Processing of fiber evaluated in this paper was completed on industrial equipment using parameters typical for 18 dpfPOY polypropylene fiber. Accelerated UV exposure testing was accomplished with xenon arc under standard interior dry xenon conditions. Sample fibers were wrapped around 6 em x 15 em cards and claluped into standard specimen holders. Tensile strengths were measured using a Instron model 1123 with a 90.7 kg (200 pound) load cell, 4 cm sample length, 12.5 em/minute pull rate and a 2.268 kg (5 pound) calibration weight.

The Influence of Co-Additive Interactions

235

RESULTS AND DISCUSSION In response to the need for the next generation light stabilizer which combines high performance, minimal interaction with co-additives, low volatility and non-extractability, HALS5 is introduced. This high performance, oligomeric HALS has a poly(methylsiloxane) backbone and is characterized by an excellent compatibility with polypropylene. As the following data will demonstrate, this uniquely designed stabilizer has perfonnance characteristics unlike any previously reported product. The soft polysiloxane backbone is quite flexible compared to traditional carbon-based products. This results in a tighter interaction with the polypropylene chain. Indeed, molecular Inodeling experiments have shown that HALS5 can adopt a helical structure along the Si-O-Si axis, with Inethyl piperidine groups spiraling along this axis. It would appear that this should have good mutual co-penetration possibilities with the helix ofpolypropylene. 2 This is consistent with the unusually high compatibility observed with polypropylene. Miscibilies in excess of 1: 1 (w/w) have been observed. This high degree of compatibility has several important consequences. High compatibility and solubility lead to excellent dispersion, probably dispersion on a molecular level. Molecular dispersion maximizes the effectiveness of light stabilizers. Other HALS, which may rely on mechanical dispersion, may not reach this level of distribution of stabilizer. The mobility of HALS5 within the polypropylene matrix is quite high. A high level of mobility generally enhances performance, except in the case ofHALSl where high mobility leads to loss ofthe stabilizer from the fiber due to volatility and a lower level of compatibility. Despite the high mobility ofHALS5, this product is nearly non-extractable. Again, compatibility tends to keep the HALS5 within the polymer matrix whereas other, less compatible materials must depend upon a high molecular weight and low rate of migration to achieve some measure of permanence. The unfortunate consequence of high molecular weight/low mobility is often low perfonnance. In designing a new stabilizer, one generally tries to maximize the percentage of active moiety. In the case of HALS, this is the tetramethylpiperdine group. Typically the backbone will only indirectly affect performance. Although this approach would appear to be quite obvious, problems such as compatibility can arise in developing prod~cts based on this consideration alone. Consequently, the ratio ofpiperdine to backbone must be balanced to arrive at the best compromise of theoretical activity and compatibility. The siloxane-based technology ofHALS5 provides a high degree ofmolar activity along with the necessary compatibility.

236

Weathering of Plastics

100

100 ~----------"""""""'-----.

%TS l(,

TS ~

HALS~

5{)

T551) • 799htS

HALS 2 ". 1$50" 431 ~f$,...

iiA:Ufj

TS50 .. 7Unt' .~~~.. .~

. 0

200

400

I 0

~-~~,,~._-----~,

()

1000

100

120

1200

"-.001\$ l~ W~M I:~:C.~~~~_~~.~ ..._......

.. Bue AJlJll! ..:::..~..l:!~_~ . ~~~~~!_ ..:,.!:iJ:'.~S2 :

Figure 1. UV stabilizing activity of HALS5 in PP-fibers exposed in WOM. Sample: PP-fiber (18 dpf).

Figure 2. UV stabilizing activity of HALS5 in PP-fibers exposed in Florida. Sample: PP-fiber (18 dpf).

% AETAIN!!D TENSIL.E STRENGTH

120·

......

PP FIBER so .

In the first set of experiments, polypropylene fibers (18 dpf) were exposed in a 40 weatherometer (bpt 63°C, RH 50%) and eval20 . uated for retention of tensile strength after o o 400 6C

60 ;

\';;

'J-

The Influence of Co-Additive Interactions

237

Table 1. Gas fading in natural fibers HALS type HALS5 HALS4

1st cvcle 4-5 4

5th cycle 4 4 )

Gray scale - (1- worst, 5 -best): Base: PP + 0.05% phosphite

'0 ~:~

1S min @ i60-C Figure 3 compares the performance of ;.z is 2ao·c several recently introduced products. In this 8asl:!. pp. O.l"!) Phenolic:PhO.s 1) .. 0 25%. HAL S .. 005% C;I$T + 0 ! B~;\o Pigment" 1 0% T102 study, HALS5 shows a distinct advantage in performance over both the new blend tech- Figure 4. HALS-pigment themla! interactions. nology product and a tertiary hindered amine, HALS4. Often the advantage of tertiary HALS is in the area of discoloration or resistance to gas fade. However, as demonstrated in Table 1, HALS5 and HALS4 exhibit essentially the same excellent gas fade characteristics in the system evaluated. Gas fade results are highly dependent upon several additional factors such as phenolic content, spin finish, catalyst residue, etc. In general, proper formulation of fiber components can often eliminate differences in the contribution of HALS to gas fade discoloration. IM'!@

PIGMENT INTERACTIONS The addition of high molecular weight HALS has been shown to affect color yield or strength of certain pigment systems. The extent ofthis appears to be related to the type of amine. Secondary amines such as HALS3 tend to have a greater propensity to affect color yield than tertiary amines such as HALS2 and HALS4. 10 In an effort to evaluate the extent of interaction of HALS5 with pigments, a thermal interaction study was carried out. In this study, the HALS and pigment were compounded and compression molded for 15 minutes at a specified temperature (260 or 280°C). The change in color was measured as delta E versus a control sample molded at 200°C for 5 minutes. Figure 4 compares HALS5 and HALS3 with their effect on color. The results are highly pigment and temperature dependent. In general, HALS5 has less negative interaction with the pigments. This is more pronounced at the higher 280°C temperature.

CHEMICAL RESISTANCE HALS are known to undergo a reaction with mineral acids such as HBr causing an inhibition of activity. For this reason, success in using HALS in combination with co-additives capable of producing acids during processing or exposure has been quite limited. Examples of

238

Weathering of Plastics

co-additives of this type are thiosynergists (DSTDP) and brominated flame retardants. Unexpectedly, HALS5 has shown an unusual chemical resistance when compared to traditional HALS. DSTDP

Thioethers such as DSTDP have been used in combination with primary antioxidants to provide extended lifetimes to Figure 5. UV stabilization ofHALS/DSTDP blends. Stretched PP polyolefins exposed to elevated temperafilms (40 ~m). Dry std. Xenon. Base: Process stabilized PP + tures. Unfortunately, the mechanism by 0.1 % DSTDPt + 0.10-0.25% HALS. which these thiosynergists function has been shown to produce sulfenic and sulfonic acids which are capable of further reacting with HALS. 11 Figure 5 shows the effect of DSTDP on HALS performance. While both HALS 1 and HALS3 show a strong negative interaction with DSTDP, HALS5 is relatively unaffected. FLAME RETARDANTS

As previously discussed, the key to HALS-flame retardant incompatibility is acid generation by the flame retardant which in tum deactivates the HALS. The mechanism for bromine radical generation by flame retardants is quite structure dependent. 12 Aliphatic brominated flame retardants are primarily decomposed thermally which may occur during the extrusion process. Altematively, aromatic brominated flame retardants are relatively stable through the processing step but may generate bromine radicals during UV exposure. Aromatic flame retardants are generally less prone to thermally induced degradation but rather generate bromine radicals through a photo-activated process. Combinations ofUV absorbers (UVA) and HALS have been shown to provide outstanding UV stability to polypropylene fiber containing aromatic flame retardants. 13 It appears that the primary benefit of the UVA is its role in providing a UV screen for the flame retardant, thus inhibiting the generation of bromine radicals and hydrobromic acid. With the level of acid minimized, even relatively basic HALS can be successfully incorporated into these formulations. The influence ofHALS structure on stabilization in these systems is examined in Figure 6. The non-basic NOR HALS (HALS6 - pKa=4.2) has a large perfOimance advantage over the basic 2 HALS3. Surprisingly, HALS5 achieved comparable stability to the NOR HALS (HALS6) despite its basic nature (pKa=9.8). Activity of this siloxane-based HALS5 can be further enhanced using a methylated analogue of the product. HALS5-Me demonstrates superior performance over that achieved by the less basic HALS2.

The Influence of Co-Additive Interactions

HOl,JRS TO T50

600 $00

/i t/ r...'" -. ~

; r · · .. ·_····· .. 400 300

~/ f

.~6§.

-395

.. .. .. .. ·.AIl·• • • · · · w ; w• • •

;laO·~i

290

rJJ._WWW.,.

:~~---~ HALS3

HALS6

...

HALS5

HAlSS·Me

All FORMULATIONS C01HAili Fill (6'l. B.). 0.5'/, tlA.l.S -1.5'1. UVA

Figure 6. Flame retardant PP fiber xenon @55°C (ASTM D-4459). All formulations contain FRI (6% Br) + 0.5% HALS + 1.5% UVA.

239 This performance attribute may be related to the unusual compatibility of HALS5 in polypropylene. A simplified rationalization of the results can be proposed. The HALS5 molecules are "content" to remain well dispersed throughout the bulk of the polypropylene matrix. Conversely, HALS3 tends to migrate towards the polar regions of sample which contain the brominated flame retardant. This unfortunately, places the HALS in direct proximity to the sites of bromine generation.

CONCLUSIONS

Hindered Amine Light Stabilizers have undergone significant advances since their initial introduction. The latest innovation is the highly compatible siloxane-based HALS. This new generation ofHALS appears to provide a significant level of improvement in its degree of interaction with coadditives such as pigments, thioesters, and flame retardants. High compatibility not only extends performance but greatly reduces extractibility. In today's era of environmental concerns, polymer permanence will continue to become an increasingly important issue.

ACKNOWLEDGMENTS The contributions and research effort of Olga Kuvshinnikova and Bill Fielding are gratefully acknowledged. Special appreciation is also extended to Dr. Carlo Neri, Dr. Patrizia Blasioli, Dr. Francesco Gratani, and the laboratories of Great Lakes Italy for their significant laboratory support.

REFERENCES 2 3 4 5 6 7

Francke, C.J.H., "Polyolefin Fibers: Growth Fibers ofthe Future". Polypropylene Technology Conference, Aug. 3 I-Sept. I, (1994), Clemson, sc. Scruggs, J.G., "A Brief Over View of the Evolution of Polypropylene Fiber Processes", Polypropylene Technology Conference, Aug. 3 I-Sept. I, (1994), Clemson, SC. Klemchuk, P. P.; Gande, M. E.; Cordola, E., PO(l'm. Deg. alld Stab., (1990),27,65. Gray, R.L., "UV Stabilization ofPP Fibers", Polypropylene Technology Conference, Sept. 4-5, (1991), Clemson, SC. Gray, R.L., "Stabilization Aspects of Additive-Pigment Interactions", Proceedings ofthe Color and Appearance RETEC, October 18-19, (1991), New Orleans, LA. Rozantsev, E.G., Free Nitroxyl Radicals, Plellum Press, New York, (1970). Dulog, L.; Bleher, H., Makromol. Gem., (1986),187,2357.

240

8 9 10 II 12 13

Weathering of Plastics

Chemela, S.; Carlsson, D.l; Wiles, n.M., Polym. Deg. and Stab., (1986), 26, 185. Neri, C.; Malatesta, V.; Ranghino, G.; Montanari, L.; Fantucci, P., Macromolecules, (1993), 26 (16),4287. Gray, R.L., "Additive-Pigment Interactions", Polypropylene Technology Conference, Sept. 4-5, (1990), Clemson, SC. Oae, S., Organic Chenlistry of Sulfur, Plenum Press, New York, (1977), 529. Gray, R.L.; Lee, R.E., "UV Stabilization Of Flame Retarded Polypropylene", ANTEC '95, May 7-11, (1995), Boston, MA. Gray, R.L.; Lee, R.E.; Sanders, B.M., "The Intluence Of Flame Retardant Structure On UV Stabilization Approaches in Polypropylene", Proceedings of the PMAD RETEC, October 18-19, (1995), White Haven, PA.

APPENDIX I HALSI

Bis (2,2,6,6-tetramethyl-4-piperidyl) sebacate, MW 481

HALS2

Dimethyl succinate polYlner with 4-hydroxy-2,2,6,6-tetramethyl-I-piperdine ethanol Mn > 2500

HALS3

Poly {[6-[( I, 1,3,3-tetramethylbutyl) imino]-1 ,3,5-triazine-2,4-diyl][2(2,2,6,6-tetramethylpiperidylamino]-hexamethylene- [4-(2,2,6,6- tetramethyl-piperidyl) -imino]}, Mn >2500

HALS4

1,3,5-Triazine-2,4,6-triatnine,N,N'"-[ I,2-ethane diyl bis [[[4.6-bis[butyl(1 ,2,2,6,6-pentamethyl-4piperidinyl) amino]-1 ,3,5-triazin-2-yl]imino]-3, I propanediyl]]-bis[N' ,N" -dibutyl-N', N" -bis( I,2,2,6,6-pentamethyl-4-piperidinyl) MW 2286

HALS5

Polymethylpropyl-3-oxy-[4(2,2,6,6-tetrmnethyl) piperidinyl] siloxane; LM: Mn 1100; HM: Mn 2200

HALS6

Decanedioic Acid, Bis{2,2,6,6-tetratnethyl-4-pipridinyl) ester, reaction products with tert-butylhydroperoxide and Octane, MW 707

New High Performance Light Stabilizer Systems for Molded-in Color TPOs: An Update

Peter Solera and Gerald Capocci Additives Division, Ciba Special!)) Che111icals, 540 White Plains Road TarJ)'folVn, NY 10591, USA

INTRODUCTION Thennoplastic olefins (TPOs) have enjoyed double digit growth in autonl0tive applications over the last several years, and the growth for TPOs over the next five years is projected to continue at a strong pace. I -2 This outstanding growth is due primarily to the excellent mechanical and physical properties possible with TPO cOlllbined with a relatively low cost. However, this strong growth is also partly due to the exceptional durability and weatherability imparted by photostabilizers. The ability to mold TPO in a wide array of colors and textures makes it an attractive nlaterial of choice for many applications for both exterior and interior automotive components. In addition, molded-in color can provide OEMs with substantial cost savings cOlupared to current basecoat/clearcoat operations. This combination of physical properties, styling flexibility and cost savings provides a large incentive for the continued growth of TPO in automotive components. However, despite the excellent weatherability of current TPO compositions, a broader adoption ofmolded-in color TPO will require superior retention of initial color and gloss over a larger color palette compared with the current fonnulations. It is this attribute, namely the retention of initial appearance of molded-in color TPOs, that is the subject of this paper. More specifically, this paper docunlents the outstanding activity of several new light stabilizer systems in molded-in color TPOs. These novel systenls were cOlupared to current high performance light stabilizer systems in pigmented EPDM and SSC (Single Site Catalysis) plastomer modified TPOs, and in a reactor grade TPO.

242

Weathering of Plastics

EXPERIMENTAL Samples discussed in this paper were produced from three basic TPO compositions and three different pigment systems. The three basic TPO compositions were as follows: 1) EPDM modified, compounded TPO containing 15% talc; 2) sse plastomer modified, compounded TPO containing 15% talc; 3) reactor grade TPO. The EPDM and plastomer modified TPOs were pigmented with either 0.25% ofPign1ent Red 3B, or 1.60/0 of a mixed, custom CPC blue/ultramarine blue pigment system. These pigments were selected based on the fact that both are known to be difficult colors to stabilize. From previous unpublished work, it was known that Pign1ent Red 3B was a difficult to stabilize colorant. As a result, in this study, formulations based on Pigment 3B provide a convenient gauge to differentiate stabilizer perfoffi1ance. The mixed blue pigment was selected because it is an actual color used in current automotive TPO applications. The reactor grade TPO was pigmented with 3% of a non-chalking grade of Ti0 2 • The compounded TPOs were prepared using conventional compounding equipment and procedures. The pigments and the various light and thennal stabilizers were added via extruder compounding followed by injection molding. Sample dimensions were 2 inches x 2 inches x 0.125 inches. Perfonnance was measured on the basis of color (~E, yellowness index) and percent gloss retention during accelerated weathering and oven aging (lower ~E and higher gloss retention meant better perfonnance). Test conditions and parameters were as follows: UV Exposure Atlas Ci65 Xenon Arc Weatherometer Exterior Automotive Xenon @ 700 e SAE J 1960 (June 1989) Gas Fade AATCC Test Method 23-1988 Atmospheric Fume Chamber 140°F Oven Aging Blue M Draft Ovens; 120°C Rotating specimen rack - 1 rpm Color Measurements ACS CS-5 Spectrophotometer Large Area View/Specular Component Included Illuminant D65/1 0 degree observer ~E*, yellowness index (ASTM D2244-79) Gloss Measurements BYK-GARDNER - Haze-Gloss Specular Gloss 60° angle ASTM 0 523-80.

243

New High Performance Light Stabilizer

RESULTS PP/EPDM

Comparison of Process Stabilizers in PP/EPDM EPDM tTIodified TPOs are typically stabilized with a phenolic/phosphite process stabilizer, a hindered amine light stabilizer, and a UV absorber. The hindered amine light stabilizer is usually a mixture of a low molecular weight and a high molecular weight HALS. The UV absorber is typically a benzotriazole. A widely used high performance light stabilizer that typifies this stabilization system is LS A (see the appendix for all stabilizer identifications). Calcium stearate is also usually added as an acid scavenger. This light stabilizer system is very effective for many automotive TPO applications and can be considered the current state-of-the-art. It was used as the control in this study. However, this high performance system was inadequate in the extremely demanding Red 3B and mixed blue pigment systems that were tested. This is shown in the following tables. Table 1. Xenon Weatherometer Exposure of Red Pigmented TPO (PP/EPDM) UV stabilizer system LSA LSB

Loadin2 level, 0.70

0/0

0.65

kJ* to onset of chalkin2 1250 4000

M: at 2500 kJ 58.4 4.2

*All weatherometer exposure intervals denoted as "kJ" refer to total incident light energy recorded as kilojoules per meter2 at 340 nanometers. All samples contain: 0.1 % calcium stearate, 15% talc. Pigment: 0.25% Pigment Red 3B.

Table 2. Xenon Weatherometer Exposure of Blue Pigmented TPO (PP/EPDM) UV stabilizer system LSA LSB

Loadin2 level, 0.70 0.65

0/0

kJ to onset of chalkin2 1920 >4000

M: at 2500 kJ 22.5 1.8

All samples contain: O. 1 % calcium stearate, 150/0 talc; Pigment: 1.6% Mixed Blue

Table 1 and Table 2 show that the stabilizer system based on the new dialkylhydroxylamine type process stabilizer was far more weatherable than a systen1 based on the traditional hindered phenol/phosphite system (i.e. LS A system). Visually, the differences in performance are even more spectacular in the red pigmented TPa compositions than what Table 1 suggests. The LS A stabilized sample exhibited severe chalking, while the sample stabilized with the hydroxylamine based system (LS B) showed only a small color change

244

Weathering of Plastics

after 2500 kilojoules of weathering, and no chalking until 4000 kilojoules. In the blue pigmented TPO compositions, the sample containing the hydroxylamine based system showed virtually no color change at 2500 kilojoules, and no evidence of chalking even after 4000 kilojoules of weathering! In comparison, the traditional LS A sample began to chalk at 1920 kJ and was severely discolored at 2500 kJ. The data also indicates that it was much more difficult to stabilize the red pigmented TPO composition than the blue TPO system.

Comparison of Light Stabilizers in PP/EPDM The previous tables showed the dramatic improvement in light stability that can be achieved by substituting a dialky1hydroxylamine process stabilizer for a hindered phenolic/phosphite based system. The weatherability of the hydroxylamine based system can be further enhanced by using more effective light stabilizers than the current state-of-the-art system. This is shown by the gloss retention results in the following graphs.

l.o.1%I.$A

.0."%1.$6

.M5%I.$C

OO.M%LSO

GO~============;::L-.. 60,----:===========~ ~ f---------.§ 50 f - - - - - - - - - ;;:: ; .ot----~~~- ~ 40 -1----------;; 'al IX 30 t - = ; : - - - - - - - t - - IX 30 . 1 - - - - - -

20 t-'=~--

f--:::-+--:--2500 kJ Xenon Weathering (SAE J 1960)

Figure I. UV stability of red pigmented IPa (PP/EPDM).

.. 20+-----~

c:; ~

.1-------

0..----2500 kJ Xenon Weathering (SAE J 1960)

Figure 2. UV stability of blue pigmented IPa (PP/EPDM).

Figures 1 and 2 show that stabilizer systems containing the dialky 1hydroxylamine were superior to the hindered phenolic/phosphite based system. In addition, the two light stabilizer systems based on HALS 3 (LS C and LS D), a new monomeric NOR type HALS, gave significantly higher gloss readings compared to the system based on the more well known HALS 1.

PP/SSC PLASTOMER Comparison of Impact Modifiers EPDM and EPR rubber have been used as impact modifiers for polypropylene for quite some time. More recently TPO suppliers have begun to use new impact modifiers that have been produced using single site constrained geometry catalysts. The weatherability of TPO compositions containing these new plastomers has not been well documented in the trade litera-

245

New High Performance Light Stabilizer

ture nor has their photostability been compared to the traditional systems. One of the objectives of this project was to compare the weatherability of PP/EPDM and pp/sse Plastomer. The following are the results of this comparison.

oor;::==============::;--

so W

c!l

.0.7 lS A wlt~ 0.25'" Fled 3B .0.65% lS 8 wit~ 0.25% Fled 38 • 0.7% lS A with I.S·;' Mixed Blue o 0.65'Y.lS B with 1.6% Mixed Blue

g 70

.. a::

~60 c; so '"

'o"

.0.7 LS A wlt~ 0.25% Fled 38 .0.65'1. LS 8

aO.6S~/.lS B

a 2O-f---

"I- '0 t-:=---" o PPlPlaslomer

2500 kJ Xenon Weathering (SAE J 1960)

Figure 3. Comparison of impact modifiers on the UV stability ofTPO.

0.25% Red 38

1.6~\

Mixed Blue

with 1.0". Mixed Blue

PPIEPDM

wlt~

.0.7% lS A with

PPIEPDM

PPlPlastomer

2500 kJ Xenon Weathering (SAE J 1960)

Figure 4. Comparison of effect of impact modifiers on the UV stability ofTPO.

Based on the reduced discoloration and the higher gloss retention, Figures 3 and 4 show that in both pigment systems the PP/Plastomer TPO compositions were considerably more weatherable than the PP/EPDM compositions. Again, the hydroxylamine based system LS B was superior to the hindered phenolic/phosphite stabilizer system (LS A), and the red compositions showed the largest shifts in color and gloss.

Comparison of Light Stabilizers in PP/Plastomer The weatherability of sse Plastomer modified TPOs can be further improved by the use of the appropriate light stabilizers. This is demonstrated in Figures 5 and 6. Figure 5 shows that the stabilizer systems based on the new hydroxylamine process stabilizer (LS B, LS C, LS D) were much better than the hindered phenolic /phosphite system (LS A) in the red pigmented TPO composition. Although there was little differentiation in hydroxylamine LS systems based on color retention examination of the gloss retention data reveals some interesting results. Figure 6 clearly shows that LS C or LS D which were based on the NOR type HALS outperformed LS B which was based on the conventional secondary amine HALS, HALS 1. Developmental High Molecular Weight Hindered Amines in PP/Plastomer As higher demands are being placed on the UV stability of plastic automotive components, the industry's requirement for new stabilizers to meet the need intensifies. Several new high molecular weight hindered amine stabilizers were examined in the most difficult to stabilize

Weathering of Plastics

246

._---_._----.~._._----

LS A

.O.MY, LS B

.0.65% LS C

OO,65,aS 0

8 0 + - - - - - - - -__

t----------

j : t----------

~ : t-~'ii:'2:""_20 ! - - T - ' ' - - 10 +--__. - -

i3 #-

3000 kJ Xenon Weathering (SAE J 1960)

()1M<. P,<:me:MR90.J.B. 0 l~<:~;:jllm S1aarllle, l!l% TalC

02;% P'.9r.'l,y.ll~¢'II-38.O.1%C~i~!l"$tIl'JJlI~. 1M1t fait;

Figure 5, UV stability of red pigmented TPO (PP/plastomer).

Figure 6, UV stability of red pigmented TPO (PP/plastomer),

pigment system (Pigment Red 313) to assess their performance against the CUITent state-of-the-art system. As previously shown, gloss retention is very often a good indication ofHALS performance. In Table 3, UV light stabilizer systems containing HALS 3 with either HALS 5 or a new developmental additive, HALS 6, outperformed the traditional HALS lIHALS 2 formulation.

Table 3. Xenon Weathering of Red Pigmented, PP/Plastomer TPO UV stabilizer svstem LSA LS B LSC LSD

Loading level, %

Gloss retention at 4000 kJ

0.70

catastrophic failure

0.65

74.0

LS E

0.65

73.8

0.65

7.1

0.65

70.8

All samples contain: O. 1 % calcium stearate, 15% talc Pigment: 0,25% Pigment Red 38

Table 3. Xenon Weathering of Red Pigmented, PP/Plastomer TPO UV stabilizer svstem

Loadinl! level, %

Gloss retention at 3000 kJ

Gloss retention at 4000 kJ

LSA LSF LSG LSH

0.70

10.9

catastrophic failure

0.70

82.0

74.9

0.70

81.9

74.0

0.70

83.3

74.7

All samples contain: 0.1 % calcium stearate, 15% talc Pigment: 0.25% Pigment Red 38

247

New High Performance Light Stabilizer

Table 4 highlights the outstanding perfonnance achieved with HALS 4, 7 and 8 in combination with HALS 3 and PS 1 compared to the control system. Gloss retentions were at least 80% of the original values at 3000 kJ - 800/0 gloss retention at 2500 kJ meets most automotive specifications. HALS 6, 7 and 8 are all high molecular weight hindered amines designed to be used in pigmented polyolefin applications requiring extended outdoor lifetimes. These new light stabilizers used in combination with the hydroxylamine process stabilizer provide an effective barrier to UV degradation. Blue pigmented plastomer/TPO with LS E, LS G, LS H yielded color shifts after 4000 kilojoules of less than 3 ~E units, and percent gloss retention values of at least 70%. This was truly an outstanding result, especially in view of the fact that the pigment system is considered to be very difficult to stabilize. 7

REACTOR GRADE TPO On a volume basis, end users consume more compounded TPO than reactor grade TPO. Compounded TPOs have a larger market share, however, reactor grade TPOs are growing at a faster rate. 8 From a weatherability and stabilization point of view, little has been published on the photostability of these polymers. A white pigmented, reactor grade TPO was included in this study as representative of this type of cOlnposition and the weathering results fronl its exposure are summarized below. As in son1e of the previous exposures, the white pigmented, reactor grade TPO samples could not be distinguished on the basis of color change (all ~Es were below 2.0). Fortunately, their UV stability could be much n10re easily assessed on the basis of gloss retention (see Table 5).

Table 5: Xenon Weathering of White Pigmented Reactor Grade TPO UV stabilizer system LSA

Loading level, 0.70

0/0

Gloss retention at 4000 kJ 8.8 6.7

LSB

0.65

LSC

0.65

66.6

LSD

0.65

68.0

Pigment: 3.0% non-chalking Ti0 2

The above results were similar to those found previously, namely that light stabilizer systems based on the use of a hydroxylamine process stabilizer and a NOR HALS were more effective than traditional systems based on a phenolic/phosphite process stabilizer and a low

248

Weathering of Plastics

molecular weight/high molecular weight NH HALS combination. Thus, the best light stability was obtained with LS C or LS D. GAS FADE DISCOLORATION In addition to providing good initial color and good UV stability, stabilizer systems must protect automotive TPOs from discoloring during warehouse storage and must protect the polymer from thermooxidation during end use. It is well known that traditional stabilizer systems based on phenolic antioxidants can lead to color formation during storage. 9 They can also form colored quinoine type compounds in functioning as antioxidants, especially at elevated temperatures. 1O It would be highly desirable if one could achieve good thermal stability and good storage stability while avoiding the undesirable color that may result from the use of phenolic antioxidants. Advanced technology, well formulated stabilizer systems can provide the above requirements and have advantages • .,_~~_,_ .....__ .."__umN_":::;:...... ..::.·:i:::..,-----· .... other than inhibition of UV degradation. •......... lSA Figure 7 illustrates the increase in resistance to discoloration on specimen plaques containing a traditional light stabilizer system - LS A - had an increase in yellowness index of more than three units after three cycles in a gas fade chamber under AATCC , Cy'¢IIfIS In Gu F.de Test Method 23 conditions. Substitution of WhhlPO .. • •_ the phenolic antioxidant with the dialkylhydroxylamine base stabilizer reFigure 7. Gas Fade Oven Discoloration - TPO (PPIEPDM). duces the likelihood of early discoloration from the presence of over-oxidized phenolic. The NOR hindered amine/ hydroxylamine combination keeps discoloration under three YI units even after seven gas fade cycles. This decrease in gas fade discoloration with LS C allows manufacturers a greater degree oftlexibility in long-term storage of molded parts compared with other stabilizer systems. In a separate study, the thermal performance of the non-phenolic based light stabilizer system (LS J) was compared to a traditional base stabilized TPO (LS I). The formulations at 120°C (Figure 8) gave roughly the same performance up to about 72 hours. At 148 hours, discoloration from the phenolic produced a LlE of7 cOiTesponding to a yellowness index ofabout 40. The hydroxylamine specimen had a color change ofless than 3 LlE

....

ft!Pl'U-e(ltlt$~

1S"'1.T~O,1't.ak:wll'l'W

Ct\a~t

249

New High Performance Light Stabilizer

....... /~~ ... ",- 0

~"'~~E'lSJ

.","" ,..+... o

.. '"'

ll<>1taE, ..

..... '•... _..... ~" ..•.. ~!:... ~._ ..

C*t;j[ ·1,.31

,",-, _..

.,,·tSJ

..6

'n,·lSI

~ ,/(c .•

.••

~<>

units. It is clear from the graph that the color shift of the white samples closely matches the change in yellowness index indicating the discoloration was due to a shift in the b* value.

•

CONCLUSIONS

A summary of the weatherability results for molded-in color TPO compositions eo '00 1:0 are as follows: 1. PS2, the new dialky 1Figure 8. TPO (PPiEPDM) - discoloration during oven aging at hydroxylamine process stabilizer, was 120°e. far more effective than the conventional hindered phenol/phosphite system (PS 1) in enhancing the light stability of molded-in color TPOs. 2. HALS 2, a relatively new monomeric NOR HALS, was far more effective than HALS 1, the traditional HALS used in light stable TPO grades. 3. For molded-in color TPOs, the most effective UV light stabilizer system consisted ofa dialkylhydroxylamine process stabilizer, a low molecular weight NOR HALS, a high molecular weight non-interactive HALS, and a hydroxyphenylbenzotriazole UV absorber. 4. LS C and LS D were much more effective than LS A and can be considered the current state-of-the-art light stabilizer system for TPOs. 5. LS C and LS D were found to be very effective in two pigment systems that were previously considered vety difficult to stabilize. 6. In gas fade testing, a system based on hydroxylamine and NOR HALS reduced discoloration by more than half compared to the traditional stabilizer package (LS A). 7. Thermal testing at both 120°C and 140°C showed that a hydroxylamine based LS system outperfonned the phenolic system in maintaining low color. 8. The high performance developmental hindered amines will continue the trend of providing end users with more effective protection systems for the industry's increasing UV stability requirements. 9. All systems containing 1.6% mixed blue pigment (Cpe blue/ultramarine blue) were found to be considerably more UV stable than those pigmented with 0.25% Red 3B. 10. With regard to impact modifiers, a plastomer produced using a stereo specific catalyst system was found to give TPO compositions that were much more weatherable than those based on EPDM. /

..... '".>

•

tb",,~Ot';lflo-41!th.11~·C

Weathering of Plastics

250

11. The significant improvement in light stability that can now be achieved with molded-in color TPOs should broaden the applications where they can adopted.

ACKNOWLEDGMENTS The authors would like to extend their appreciation to Noe Castillo and James Osmundsen who carried out the sample preparation and testing for these studies. The authors would also like to thank the management of Ciba Specialty Chemicals Corporation for permission to publish this paper.

APPENDIX I IDENTIFICATION OF STABILIZERS AND STABILIZER SYSTEMS STABILIZERS

PS 1 PS 2 HALSI HALS2 HALS3 HALS4 HALS5 HALS6 HALS7 HALS8 UVA 1

1/1 blend of Irganox 1010 and lrgafos 168 == Irganox B 225 FS 042 == Dialkylhydroxylamine Tinuvin 770 Chimassorb 944 Tinuvin 123 Chimassorb 119 CGL 116 Developmental HALS with a proprietary structure Developmental HALS with a proprietary structure Developmental HALS with a proprietary structure Tinuvin 328 LIGHT STABILIZER FORMULATIONS

LS A LS B LS C LS D LS E LS F

O. 1 % Irganox B 225/0.2% Tinuvin 770/0.2% Chimassorb 944/0.2% Tinuvin 328 0.05% FS 042/0.2% Tinuvin 770/0.20/0 Chimassorb 944/0.2% Tinuvin 328 0.05% FS 042/0.2% Tinuvin 123/0.20/0 Chimassorb 119/0.2% Tinuvin 328 0.05% FS 042/0.2% Tinuvin 123/0.2% CGL 116/0.2% Tinuvin 328 0.05 FS 042/0.2% Tinuvin 123/0.20/0 HALS 6/0.2% Tinuvin 328 O. 1 % Irganox B 225/0.2% Tinuvin 123/0.2% Chimassorb 119/0.2% Tinuvin 328

251

New High Performance Light Stabilizer

LSG LSH

LS I LS J

O. 1 % Irganox B 225/0.2% Tinuvin 123/0.2% HALS 7/0.2% Tinuvin 328 0.1 % Irganox B 225/0.2% Tinuvin 123/0.2% HALS 8/0.2% Tinuvin 328 0.2% Irganox B 225/0.2% Tinuvin 770/0.2% Chimassorb 944/0.2% Tinuvin 328 0.1 % FS 042/0.15% Tinuvin 770/0.1 % Chimassorb 119/0.15% Chin1assorb 944/0.2% Tinuvin 328

PS 1

PS2

<~ L~ r\-t~

/ \

b,p/o-

1:)

HO--N

C1.H 31·n

C 1.H 37·n

FS042

Irgafos 168 (50%)

~oII 'I

CHrCH;--C-O-CHt'kC

Irganox 1010 (50%) HALS 1

H{=>-o.tICH".t0-QH

HALS2

Tinuvin 770

pi P.-Q

HAlS3

O~C-lCHtltC'O

C.H;tO-N

W-~~-l

Chimassorb 944 HALS4

N~O-C.H,;

Jl1j'",yR NyN

'''It''·~(tl~ . . . . .

NiN

Tinuvin 123

N'' ____

'l,

"'-r-c}-CH, C.H.

HALS5 H-N

H.e., -d f

.... R

R:::;H,

HALS8

Developmental HALS

~~·c

N-t

R-N R

<~~,

HALS6

Chimassorb 119 Developmental HAlS

H/~J

N-R

:N",r.J

•

CGL-116 HALS7 UVA 1

Developmental HALS

qCICH,',CH,cH,

-, 0:"...

--N 1

N- ~

\!J tACH.),CHFH.

.............

252

Weathering of Plastics

REFERENCES 1 2 3 4 5

6 7 8 9 10

R. Price, "Outlook for TPO Resins in Automotive Applications", TPOs in Automotive, October 17-18, 1994. D. Blank, C. Buehler, M. Paschick, "High Gloss TPO Materials With In1proved Durability", TPOs In Auton10tive 96, October 28-30, 1996. E. Lau, D. Edge, "Novel Precolored TPO Systems for Partially Painted and Non-Painted Exterior Automotive Applications", Annual Technical Conference of the Society of Plastics Engineers, May 9-13, 1993. D. Blank, C. Buehle~, M. Paschick, "High Gloss TPO Materials With Improved Durability", TPOs In Automotive 96, October 28-30,1996. 1. Dibbern, M. Laughner, H. Silvis, "Polypropylene Modification With Elastomeric Ethylene/Octene Copolymers Produced by Single Site Constrained Geometry Catalyst", SPE Polyolefins X International Conference, February 23-26, 1997. T. Yu, "Plastomer-Polypropylene Blend Mixology", SPE Polyolefins X International Conference, February 23-26, 1997. F. Rodrigues, "Effect of Weathering on Ultramarine Blue Pigment in Polyolefins", Annual Technical Conference of the Society of Plastics Engineers, May 3-7, 1992. R. Eller, "Business and Teclmical Trends in Automotive Applications", TPOs In Automotive 96, October 28-30,1996. Smeltz, K.C.; 'Why Do White Fabrics and Garments Tum Yellow During Storage in Polyethylene Bags and Wrappings?"; Textile Chemist and Colorist, Vol. 15, No.4; pp 52-56; 1983. Klemchuk, P. and Homg, P. L., Polymer Degradation and Stability, 34; pp. 333 - 346 1991.

Stabilization of Polyolefins by Photoreactive Light Stabilizers

Gilbert Ligner and Jan Malik Clariant Huningue S.A., F-68331 Huningue Cedex, France

INTRODUCTION The parameters most etnphasized in any consideration ofphysical aspects ofpolymer stabilization are diffusion, solubility and volatility of additives. It is generally accepted that an efficient additive should be well soluble in the polymer to be stabilized, whereas views on the importance of the mobility or diffusiveness of additives are not so firmly established. Many authors suggest that the efficient stabilizer should be able to diffuse easily throughout the polymer matrix. However, the diffusion rate should be optimal i.e. the stabilizer should be mobile, but not so mobile that appreciable quantities are lost. Several studies l ,2 have attempted either to estimate the optimal molecular weight of an additive or to give an explanation in terms of lower compatibility and lower mobility for the observed decrease of the stabilization efficiency brought about by increasing n10lecular weight of the additive. 2-6 Unfortunately, neither consistent studies of molecular weight nor concrete data on compatibility or mobility of stabilizers have been reported. A more probable explanation has been offered in terms of the hom*ogeneity of distribution of active stabilizing moieties. Polymerized additives with high molecular weight exhibit inhom*ogeneities in the distribution ofpiperidine active sites, leaving a great pat1 ofthe polyolefin matrix unprotected. 7-12 Contrary to the concept of"optimal mobility", Moisan 13 reported an empirical relationship based on his experimental findings. According to Moisan's data, the efficiency (ter) depends on the solubility (S) and diffusion coefficient (D) as defined by the following relationship: ter= f [In(S2/D)]. This n1eans that an effective stabilizer should be highly soluble in the polymer and its diffusion rate should be minimal. Based on this concept, new developments have been oriented towards polymer-bound stabilizers. A reactive processing technology has been proposed for several years,14-17 the principle of which is that a "reactive HALS" is melt-processed with a polyolefin in the presence of peroxide, resulting in the attachment of the HALS molecule to

254

Weathering of Plastics

the polymer. More recently, a new type ofmolecule based on a benzylidene malonate structure has been specially designed to become R1-o-CH=C c-o II grafted to polymers by a photochemical reaco 17 20 2 tion. The basic structure is given by HALS-4 HALS-4. HALS-4 represents the results of intensive research on the photochen1istry of HALS secondary structures. The principle of action of this type of product is the production of a "reservoir effect", as found with HALS-l type compounds (migration of the low molecular weight material to the surface), followed by a photoreaction of the product with the polymer at the surface (thus giving the long-lasting effect characteristic ofHALS-2 and HALS-3 materials).

PHOTOREACTION AND PHOTOGRAFTING MECHANISMS In previous work, attempts have Table 1. Time for total photoreaction in model been made to systems phenomenologically describe the photoreaction which is expected lVloJecule to take place between the T,h Rl R2 designation methylenic double bond of 6 HALS-4/1 H H HALS 4 and the host matrix by HALS-4/2 3.5 H Me means of photo-initiated radiHALS-4/3 MeO H 12.5 cals. 19-22 It has been shown that HALS-4/4 MeO Me 9 the photoreaction can be monitored by the decrease of the UV absorption of the product. In the present study, various investigations on the photoreactivity of the HALS-4 type were carried out. Model solutions containing 0.5% of HALS-4 analogs were exposed to UV light (high pressure mercury lamps with cut-off filter for wavelengths below 290 run). The study was carried out with molecules having different Rl and R2 substituents. The time to total photoreaction (T) of various analogs was measured as the time to total disappearance ofUV absorption in THF solutions. The results of the n10st interesting molecules are reported in Table 1. It is seen that the presence ofa methoxy group as RI instead ofH as well the presence of a methyl group as R2 decreases the reaction rate significantly. The same observation was made with the additives incorporated in the polymer matrix. The HALS-4 type products are

Photoreactive Light Stabilizers

255

able to provide a reservoir effect without blooming only if the rate of photografting is higher than the rate of migration. As seen above, the rate of photografting can be controlled by the Rl and R2 groups so that the light stabilization performance ofHALS-4 can be enhanced by choosing the right substituents. Products such as HALS-4/l and HALS-4/2 have too high a photografting rate, which limits the reservoir effect (rapid total grafting). On the contrary, products such as HALS-4/3 have too Iowa photografting rate, probably causing wash-out ofa part of the product during weathering. It was found that HALS-4/4 meets the requirement for an optimal efficiency. In fact, HALS-4/4 is the first and hitherto the only commercialized product of this type. The influence of Rl and R2 on the photochemistry was investigated with model solutions as well as in polymers. GC-MS and NMR measurements suggested that the mechanism is rather complex. It can be expected that the grafting could take place by a one-step reaction of polypropylene with the additive in an excited state and/or by a two-step reaction involving radical intermediates. This second mechanism could give some understanding of the influence of the substituents Rl and R2, Ab.o,b.nee 1% of Inltl.!) C.,bonvl Index e.g., the induction effect of a 120....--.,.--------..---------. methoxy group in Rl position could De '.d"ion stabilize the radical intermediate. 100 0.6 It should be pointed out that the 80 0.6 action of the piperidine group as a radical scavenger is not influenced 80 0.4 by the grafting mechanism which is 40 described in the previous sections. 0.2 As in the case of all "conventional" 20 HALS, the light stabilization is o -l-~~=----==----=T---=:::::==~....---.,.....JO clearly observed when the polymer is o 500 1.000 1.500 exposed to UV light. This can be eviTim. of hp<>.ur. IH<>uro) denced by FTIR measurement. Figure I. Spectroscopical measurements of exposed PP containing Carlsson and Wiles 23 showed that the $labi~u1iQn

photoreactive HALS.

consumption of a HALS can be monitored by measuring the carbonyl absorption of a carbonyl-containing additive by FTIR spectroscopy as a function ofthe exposure time. Figure 1 shows the curves obtained with HALS-4/4 in polypropylene. The IR absorption at 1736 cm- I con'esponds to the carbonyl group of the additive. It is seen by the decrease of the curve that the light stabilizer acts immediately. In addition, the grafting period (G) which can be assessed by the UV absorption decrease (or IR absorption at 1724 em-I) overlaps the stabilization period. As usual, degradation is obtained from the absorption at l7l5cm- l .

256

Weathering of Plastics

MODULATION OF THE GRAFTING RATE INFLUENCE OF WEATt-IERING CONDITIONS In previous sections, it has been

Table 2. Time to failure of HALS stabilized PP shown that the photografting rate films can be influenced by nl0lecular design. In any case, grafting process starts as soon as the polymer sur0.1% HALS-2 1440 hrs 1430 hrs 23 months face is exposed to light. The 1650 hrs 2200 hrs 33 months 0.1 % HALS-4/4 photoreaction ofHALS-4/4 can be 0.075~~ HALS-4/4 1250 hrs 28 months 2050 hrs completed in a polypropylene (PP) film after 200 to 300 hours expoPre-exp.: 3 weeks pre-exposure followed by Weather-Ometer (WOM) sure to a Xenon lamp exposure. (Weather-Ometer under dry conditions).18 In the case of natural exposure, the photoreaction clearly depends on season and the sunshine exposure; grafting periods vary between two weeks in summer and forty days in winter. In any case, optimal conditions, in tenns of the kinetics of grafting and migration, are obtained by natural exposure. 21 ,22 However, many testing methods involve accelerated weathering devices. Thus, a realistic evaluation of perfonnance of HALS-4/4 compared to "conventional" HALS can be obtained by pre-exposing the stabilized polymer to natural conditions until the grafting is completed. The second step can be achieved by artificial weathering. Experiments were carried out with 100 microns thick PP films. Results are given in Table 2. The results of Table 2 indicate that even a lower concentration of HALS-4/4 is able to achieve a level of stabilization similar to that achievable by use of a typical concentration of the oligomeric HALS-2. It should be noted here that the photoreactive HALS-4/4 is also photosensitive to spectral distribution of light. Unadapted artificial light such as high energy emission (UV-B) can lead to partial destruction ofthe additive with only limited grafting. It is seen once more that artificial weathering does not always reflect reality since it does not correlate with natural exposure. Thus, the real advantages of the photografting HALS-4/4 which are obvious in natural weathering can be observed in accelerated weathering if the testing conditions are properly adopted to simulate outdoor exposure. w/o Pre-expo

with Pre-expo

Natural test

SCREENING EFFECTS In thick PP articles exposed to light, the grafting reaction takes place froln the surface down to approximately 100 microns in the same time scale as those observed with films. This can be evidenced by microtoming polymer plaques after light exposure and by analyzing each slice.

257

Photoreactive Light Stabilizers

A "free additive" concentration profile can be obtained by this method?1 Very similar profiles can be obtained by exposing film stacks. However, this second • •, lrllllr. UV method makes things much easier, since 114 liIrnt the films can be rapidly separated for AlmVV analysis. It also permits the determinaC: 1 film tion of the influence of various parame--.",.".. ".. _ ters separately such as the internal screening effect of the polymer. Figure 2 +Sun light shows the "grafting profiles" which are obtained with PP film stacks. Six films Film Numb.r containing HALS-4/4 were screened by Figure 2. Screening effect on the grafting reaction observed with PP films stacks in outdoor exposure. various additi ve- free polymer films. The total stacks were exTable 3. Q-UV UV-A exposure of LOPE press films posed in natural conditions. It can clearly be Additive formulation Measured value*, h Calculated value, h seen that the presence No HALS 423 of the additive-free 0.3% HALS 4554 polymer films delays 0.3% UVA-I \165 the grafting. This ef0.3% UVA-2 776 fect is obviously more 0.1 % HALS-4/4 + 0.2 % UVA-I 4730 2573 pronoun- ced in the 0.2% HALS-4/4 + 0.1 % UVA-I 5295 3883 presence of pigments 0.1% HAL- 4/4 + 0.2 % UVA-2 3847 2502 or UV absorbers. It 0.2% HALS-4/4 + 0.1 % UVA-2 4165 3741 was found that the screening effect can be very beneficial for the light stabilization performance of HALS-4/4 as documented in Table 3. The calculated values represents the "additive effect", i.e. the mathematical addition of the stabilizing contribution of single additives obtained from stabilization performance data of the single additives. It can be seen that combinations such as 0.2% HALS-4/4 particularly with 0.1 % UVA-l, permit a significant improvement in the light stabilization ofLDPE. Similar results were obtained with Xenon arc exposure (WOM). R.malnlng "fr•• HAlS (% of lnilioll

258

Weathering of Plastics

PIGMENTED SYSTEMS Combination of HALS-4/4 with pigment are potentially of similar interest to con1binations with UV absorbers. It was found that HALS-4/4 shows much lower influence on pigments after processing than HALS-2 does. Table 4 shows results obtained by colorimetry of 1.5 mm injection-molded PP plaques containing a yellow azo pigment.

Table 4. Color change of yellow pigmented PP plaques induced by HALS addition Control

0.25% HALS-2

0.5% HALS-2

0.25°A. HALS-4/4

0.5% HALS-4/4

0.2

-0.2

0.8

1.0

delta C

-1.6

-1.8

delta E

2.5

2.3

The delta C and delta E values were measured against a control which contained no HALS. The results show that, after incorporation of HALS-4/4, there are already considerably smaller deviations fronl the desired shade than with plaques containing HALS-2. After 600 hours exposure to Q-UV UV-A, new measurements were made, and the values reported in Table 5.

Table 5. Color change of yellow pigmented PP plaques after weathering Control

0.25% HALS-2

0.5% HALS-2

0.250/0 HALS-4/4

0.5% HALS-4/4

delta C

-1.3

-0.6

-0.7

delta E

1.7

2.1

-1.2 1.5

1.1

1.2

Delta C and delta E correspond to the color change due to weathering. The control shows the color change which is obtained without HALS. Comparison of the test results clearly shows the color stabilization activity of HALS-4/4. While HALS-2 has almost no effect on the light fastness of the pigmentation, an improvement is obtained by using HALS-4/4 ; i.e. the measured color deviation due to light exposure is substantially reduced compared to the original state. In the present case, a concentration of0.25% ofHALS-4/4 is already sufficient. A further example which illustrates the color stabilization ofHALS-4/4 is given by test results obtained from Ziegler-type HDPE injection molding plaques containing "ultramarin blue" pigment. Color values after 9000 hours NATAC 200 (natural accelerated weathering carried out in south of France) are reported in Table 6.

259

Photoreactive Light Stabilizers

Table 6. Color change of blue pigmented The presented results show that the HOPE plaques after weathering photoreactive HALS 4/4 is very efficient in PP and HDPE. According to the latest 0.15% HALS-2 0.15% HALS-4/4 indications, this type of HALS n10lecule 9.1 5.0 delta E also performs very well in rubber-modified and/or filled polymers as well as in styrene polymers, PBT and

pvc. CONCLUSIONS Most of the works dealing with HALS performance emphasize the importance of cOlTIpatibility and solubility of the additive. Terms such as "reduced migration" or "limited compatibility" are used very often to explain differences in the stabilization performance, especially in connection with oligomeric and polymeric additives. The available physical measurements, as well as their relation to the stabilization efficiency, imply three empirical requirements for an effective stabilizer: high solubility, n1inimal diffusion, and high hom*ogeneity in the distribution of active species. The data presented in this paper shows that, under optimal conditions, the use of photografting HALS-4/4 could satisfy all three en1pirical requirements. As a low molecular weight HALS, this stabilizer is readily hom*ogeneously distributed in the polymer during the processing step. The chemical structure ofHALS-4/4Ieads to expectation that its behavior in a polymer matrix (in terms of diffusion and solubility) would be similar to other low n10lecular weight stabilizers. Subsequent light exposure initiates a photochemical reaction by which the additive can be grafted to polymer chains. Once the additive is chemically bound to the polymer, its diffusion is minimized while the solubility is increased since the bound stabilizer has become a part of the polymer chain. The reported results obtained so far with this product confirm the expectations and show a better stabilization performance that of conventional low lTIolecular weight and oligomeric additives.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

F. GugmTIus: Research Disclosure, 209, 357 (1981).

M. Minagawa: Polym.Deg.& Stab., 25 (1989),121. P. Hrdlovic, S. Chmela: Vybrane problemy stabilizacie monomerov a polymerov, Preprints, Bratislava 1986, pAl. S. Chmela, P. Hrdlovic: 11 th Discussion Conference on Chemical and Physical Phenomena in the Ageing of Polymers, Prague 1988, P9. S. Chmela, P. Hrdlovic, Z. Manasek: Polym.Deg.& Stab., 11 (1985),233. S. Chmela, P. Hrdlovic, Po~vm. Deg. Stab., 11 (1985),339. J. Malik, A. Hrivik, E. TOlllova, Polym. Deg. Stab., 35 (1992), 61. J. Malik, A. Hrivik, D. Alexyova, Po(vm. Deg. Stab., 35 (1992), 125. J. Mallik, A. Hrivik, D. Q. Tuan, P. Alexy, P. Danko, Polymer Preprints, Vo1.34, No.2 (1993), 170. V. Dudler, Po(rm. Deg. Stab., 42 (1993), 205.

260

11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

Weathering of Plastics

J. Pospisil, Advances in Polymer Science, Vol. 124, Springer Verlag Berlin Heidelberg 1995, p.87-189. N. C. Billingham & P. D. Calvert, in Development in Polymer Stabilisation - 3, Ed. G. Scott, Applied Science Publishers, London, 1980 p. 139. J. Y. Moisan, Polynler Permeability, ed. J. COlnyn, Elsevier, 1985, 119. S. AI-Malaika, A. Q. Ibrahiln, G. Scott: Polym.Deg.& Stab., 22 (1988), 233. S. AI-Malaika, Polym. Plas. Techn. Eng., 29 (1990),73. S. AI-Malaika, Chemtech, June 1990, 366. J. Malik, G. Ligner and L. Avar, Lucerne 96. G. Ligner, L. Avar, Polymer Preprints, Vol. 34, No.2 (1993), 160. G. Ligner, Conf. Satcar '95, Clermont-Ferrand (France), May 17-18 1995. C. Werrion and G. Ligner, lIe Joumees d'Etude sur Ie Vieillissem*nt des Polymeres, Bandol (France) september 20-22, 1995. G. Ligner & 1. Malik, Conference on Additives for Metallocene Catalyzed Polymers, June 25-26, 1996, Chicago III. G. Ligner & 1. Malik, ADDCON 96, May 21-22, 1996 in Bruxelles (Belgium) D.1. Carlsson and D. M. Wiles, J. Alacromol. Sci.-Rev. Macromol. Chem. C, 14(2) (1976) p. 155.

Effect of Stabilizer on Photo-Degradation Depth Profile

T. J. Turton and J. R. White Universit)J ofNewcastle upon Tyne, UK

INTRODUCTION Ultraviolet (UV) photo-oxidation is a common cause of polymer degradation. In hot sunny climates the reaction with parts made from polyethylene, polypropylene and many other polymers is so rapid that the oxygen is consumed before it can penetrate far into the interior of an unstabilized part and degradation is confined to a region near to the surface. This has been reviewed by Audouin et al. l Some recovery of oxygen levels can occur in the interior during the hours of darkness but a sharp gradient of degradation develops in most thick parts (thickness ~'"'-'2 mm). This is expected to be even more pronounced in laboratory tests in which the UV irradiation is applied 24 hours per day. The extent ofdegradation can be assessed by infrared analysis (e.g., following the build up of carbonyl groups or other products of oxidation as by Fumeaux et al. 2 ) or by molecular size analysis (e.g., as by White and co-workers 3-5) using samples extracted at different depths from the surface. When the polymer contains a photo-stabilizer the oxidation rate is much reduced. Because ofthis the oxygen is not consumed so completely. Therefore oxygen is readily available at quite large depths from the surface of thick polymer parts if they contain an effective photo-stabilizer. Stabilizers reduce but do not completely prevent oxidation so it can be expected that some reaction will take place in the interior of a stabilized polymer. Thus in the interior of a thick part made from a stabilized polymer the degradation may be greater than in an unstabilized part when conditions favor rapid reaction (especially for high intensity UV applied uninterrupted). The studies reported here were aimed at verifying this in a family of polypropylene injection molding materials.

262

Weathering of Plastics

EXPERIMENTAL MATERIALS AND SAMPLE PREPARATION The materials used in this study were based on Montell Moplen polypropylene grade EPF 30U which was provided in both unstabilized and stabilized fonn. The stabilizer system was formulated fron1 the Ciba-Geigy range and consisted of 0.3% Tinuvin 770, a monomeric hindered amine light stabilizer (HALS), 0.3% Chimassorb 944 (an oligomeric HALS), and 0.3% Irganox B215 (a phenolic anti-oxidant). Moplen EPF 30U is toughened by a rubbery ethylene-propylene copolymer which separates as spherical inclusions. A second similar PP made by a different route but possessing many properties similar to that ofEPF 30U was supplied by the manufacturer as X-EPF 30U (=MPP in the coding used below). It was provided in both unstabilized (MPP) and stabilized (MPPS) form; pigmented X-EPF 30U was also provided in both unstabilized (MPPP) and stabilized (MPPSP) fonn. The stabilizer system was the same as that used with EPF 30U and the pigment was rutile Ti0 2 (1 %). Tensile test bars measuring ~ 192 mm x 12.7 mm x 3 mm were injection molded using a single end-gated cavity. EXPOSURE CONDITIONS Samples were exposed on open racks using Q-Panel UVA-340 tubes with output in the UV matching the solar radiation spectrum at the Earth's surface fairly closely in the wavelength range below 360 run down to the cut-off at approximately 295 nm. 6-8 This has been verified by measurements of the spectral output of the UVA-340 tubes made using a Bentham Instruments spectroradiometer based on a double grating monochromator. 7 In tests in which continuous exposure was applied (24 hours per day) the intensity at the san1ple surface was 2.0-2.3 Wm- 2 in the wavelength range 295-320 nrn, that is the total radiation below 320 nm wavelength. Some tests were conducted using a shutter that gave 12 hours onll2 hours off- in the cyclic exposure rig the intensity at the sanlple surface was 2.4-2.6Wnl-2 in the wavelength range 295-320 run. The illumination was checked regularly using the spectroradiometer and tubes were changed as necessary to maintain the desired level. CHARACTERIZATION Exposed bars were renl0ved periodically for testing. Samples were extracted at different depths from the surface by high speed milling using a single point cutter with fly cutting action, a method shown in previous studies to be suitable. The machine bed was thoroughly cleaned before machining cotnl11enced. Each milling pass removed a depth of 0.1 mm and the material removed was collected for analysis. Each sample was labelled according to the depth

Photo-Degradation Depth Profile

263

of the mid-plane, for example the third machining pass removed n1aterial between 0.2 mm and 0.3 mm from the original surface and was labelled as coming from a depth of0.25 mm. Chemical degradation was assessed using gel permeation chromatography (GPC). GPC measuren1ents were made at RAPRA Technology Ltd., Shrewsbury, UK using procedures described elsewhere. 3 After analysis the results were displayed as molecular weight distributions and the number average molecular weight (M n) and the weight average molecular weight (Mw ) were computed for each sample.

RESULTS MOLAR MASS DISTRIBUTIONS Molar mass distributions obtained from samples extracted at different depths after 49 weeks UV exposure are shown in Figures 1 and 2 for unstabilized EPF 30U and in Figures 3 and 4 for stabilized EPF 30U. Figure 1 shows the results obtained with samples taken froin near the exposed surface ofunstabilized EPF 30U and from a reference sample extracted from a virgin granule (unprocessed, unexposed). The molecular size distribution for the sample taken from the top 0.1 mm ("0.05 mm" sample) has shifted to the left from the reference sample distribution by more than one decade, indicating multiple scission had taken place. The shift towards smaller molecular weights is progressively smaller for samples taken from locations deeper within the bar but is still very significant for the 0.35 mm sample. For this one there is a high molecular weight tail that is similar to that obtained with the reference sample. A n10re detailed analysis confirms that this corresponds to some limited crosslinking that continues alongside the predominant scission event. 9 On continuing deeper into the bar it was found that the shift in the molecular size distribution was much less than with samples taken from near to the surface (Figure 2). It is notable that the high molecular weight tail for samples taken from near to the center ofthe bar lies above that for the reference sample and is evidence for significant crosslinking in this region. Figure 3 shows molecular size distributions for samples taken from within the top 0.4 mm from the surface of stabilized EPF 30U. There is a small but significant shift from the reference profile to smaller molecular weights for all samples but there is not very much difference from one degraded sample to another. The distributions from deeper locations shown in Figure 4 are fairly similar showing that the molecular degradation is nearly the same at all depths. There is no significant indication of crosslinking in the molecular size distributions from degraded sainpies in either Figures 3 or 4. It is of particular interest to con1pare the distributions for samples taken from locations in the interior of unstabilized (Figure 2) and stabilized (Figure 4) bars. It is evident that there is a greater shift to smaller molecular size in the stabilized bars. There is greater evidence for crosslinking in the unstabilized bars.

Weathering of Plastics

264 1.0 tH'i\lIb,il:~d

49 1Il'~.U .l~Oiurt

0.8

Q.a

0,6

g'" ~.

0.5

g'"

0.<

0.2

f 0.'

0.2

!og M

log M

Figure 2. Molecular size distribution for unstabilized PP at different depths in the interior after 49 weeks exposure.

Figure I. Molecular size distribution for unstabilized PP at different depths near the surface after 49 weeks exposure.

1.0 , - - - - - , . . . - - - . . . - - -........- - - - , . - - - . . ,

" ""

r~ ~1abd,Ud pp <9 wee"

,,~..,"

.. .I'j //.....

~;,

; >' 0.' ..

l I

"',.1,' ~.iI

0.\,02",,,, 02-0 j",,,, -',03-04",,,,

)17 ' \\ ,'.If

0.8

~ _ 0.5

0-0.1"''''

J \ ' "---

"'0

.- •••

'J \

'\\ \ \'. \

.. e'~r

.'''biliud PP

r~ftr"net

'\~"

;;>

0.'

0.2

log M

- - - O.S-O.&mm

·O,,]'-O,Sr.":m

0S I0

\1. \

!,!-I.lmm

Ii. \

"V.,

f/; /

I a

·····O.«-O.Smm

/t/ '" , -.""' I'

.;:0'

1J ,,:

,:J!\. I ~

/,..1/\'\ i ~

49 ••• k~ UPOlut I

,t. .

'\'. \

'--__"'.....~, ''"--

~~/...

_--=:":;';~=_-'-

1....

'\. \

~":::'_...J

lag M

Figure 3. Molecular size distribution for stabilized PP at different depths near the surface after 49 weeks exposure.

Figure 4. Molecular size distribution for stabilized PP at different depths in the interior after 49 weeks exposure.

MOLECULAR WEIGHT AVERAGES: EFFECT OF EXPOSURE TIME Data ofthe kind displayed in Figures 1-4 were obtained for a range of exposure times up to 64 weeks. Although the fine details of molecular degradation require consideration of the full size distribution, it is often more convenient to use one of the molecular weight averages to follow the progress of degradation. Figure 5 shows the variation of M w with depth in unstabilized EFP 30U for exposure times of 16, 34,49, and 64 weeks. It is evident that the major changes take place within 0.5 mm of the surface. The behavior is quite different with the stabilized grade (Figure 5) which shows progressive reduction in molecular weight with in-

265

Photo-Degradation Depth Profile

distance from exposed surface

(mm)

Figure 5. M,\ versus depth for unstabilized PP san1ples after exposures of 16,34,49, and 64 weeks.

distance from exposed surface

(mm)

Figure 6. M,\ versus depth for stabilized PP samples after exposures of 16,34,49, and 64 weeks.

creasing exposure time but very little variation in Mw through the depth for any chosen exposure. EFFECT OF PIGMENT

Figure 6 shows the variation of Mw with depth after 16 weeks exposure for a series of polymers based on the experimental grade ofPP: X-EPF 30U. Prior to degradation this polYlTIer was found to have a value ofM w approximately 25% higher than that of the EPF 30U. The unstabilized base polymer showed a variation in M w with depth rather similar to that shown in Figure 5 for EPF 30U after the same exposure. The presence of stabilizer had the same effect as with EPF 30U (compare Figures 5 and 6). The presence ofpigment caused the degradation to fall to minimal levels within 0.3 mm of the surface in both the unstabilized and the stabilized PP (Figure 7). Even in the unstabilized polymer, degradation was very restricted even at 0.1 mm depth. After longer exposures the steep part of the M w versus depth curve for unstabilized unpigmented PP moved to the right whereas for stabilized unpigmented X-EPF 30U the M w versus depth relationship remained quite flat and lTIoved down progressively with increasing exposure, similar to the behavior shown in Figures 5 and 6 for EPF 30U. EFFECT OF ON/OFF UV CYCLING

Figure 8 shows the variation ofMw with depth for unstabilized EPF 30U for a bar exposed cyclically to UV (12 hours onJ12 hours off) for 32 days compared to that obtained after 16

266

Weathering of Plastics 300

o o

::i

X-EPF 3CU

cOI\'l~o~ne,:

2S0.,r. r v.: ....... 'f""'V-b~,---.----------...... ti;~~ 200 ~ t

ISO

100

fe-."O-"O.

..-..

~)'

=:::s

• - - - - MPPP

of o

4""""""~~~

~...

•

..,

l '

0.5

MPPSP ,

~ 1

150

1 1

g --...

•

1 !

1.0

'1.5

tmm}

Figure 7. Mw versus depth for PP compounds after 16 \veeks exposure.

. . . . . . ---__,

....-.'"'l_-..,..----r----~~

tt--~-----------rtt/ r

,.:'P! -e••.•...•..••..••..•••

100

distance from exposed surf ace

200

16 week. upo.Uft

:I .' I ./i? ~

~nsubliiz.d

• <J

~ ~

MP?

•- - ';8 weef(s unlntetrupted ~~ 0

weeit.s ctl/Qff eyc:lu

0"----------:---------------. .... LO o

1.5

distance from exposed surface

(mm)

Figure 8. M w versus depth for unstabilized PP samples after 16 weeks continuous exposure and after 32 weeks of cyclic exposure (12 hours cycle).

weeks continuous exposure (i.e. having approximately the same accumulated UV exposure). There are clearly significant differences between the two sets of results but at this stage it is possible only to speculate on the interpretation (see below). The nlolecular size reduction near the surface (~O.5 nun from the surface) is greater in the bar exposed continuously whereas the size reduction nearer to the center of the molding was greater in the bar that was exposed cyclically.

DISCUSSION Figure 5 shows that a steep degradation profile had already developed in unstabilized PP by the time the first samples were removed (16 weeks exposure). The molecular size near the surface had fallen significantly by this time indicating several scissions per molecule. M w , the weight average molecular weight, had also fallen in the interior. Continued exposure caused further molecular weight reduction near to the surface; a powdery residue formed on the surface that was not analyzed and which presumably contained very small molecular weight fragments. M w appeared to recover in the interior on continued exposure. We speculate that at the beginning ofthe exposure scission was the predominant molecular degradation effect at all depths. Reaction in the interior at this time was allowed by the pre-existing equilibrium concentration of oxygen. Reaction near the surface can be sustained by oxygen diffusing in from the surrounding air but it is consumed before it can penetrate far into the bar. The reaction in the interior then continues at very low oxygen levels causing a reduction in rate and changing its character.

Photo-Degradation Depth Profile

267

This favors crosslinking reactions between long chain radicals rather than further reactions involving oxygen (see references 10,11). Thus M w increased slowly until after 64 weeks it had almost returned to the original value. The molecular weight distributions shown in Figures 1 and 2 support this interpretation. Figure 1 shows that scission dominates near to the surface. The molecular weight distribution from the sample from 0.3-0.4 mm depth is shifted considerably to the left of the distribution fronl the reference sample, indicating a large amount of scission, but shows a high molecular weight tail that suggests some crosslinking has occurred as well. In Figure 2 the high molecular weight tails for samples taken from the deepest sites lie above the reference distribution in this region (logM>6); this seems to be clear evidence for crosslinking. Thus even though the molecular weight averages for the material in the interior after 49 weeks exposure are not very different from the undegraded (reference) value, the molecular weight distribution is different to that for the undegraded material and it is deduced that SOllle reaction has occurred. In the stabilized grade the average molecular weight fell progressively with exposure (Figure 6). No high molecular weight tail is evident at any depth with this lllaterial (Figures 3 and 4). The whole molecular weight distribution shifted to the left at all depths. The amount of shift was quite similar at all depths. Near the surface of the sample the fall in molecular weight was much less than that observed in the unstabilized samples, indicating that the stabilizer system is very effective in reducing chain scission. In the interior, the fall in average molecular weight was greater than that observed in unstabilized material. This is believed to be the consequence of continued availability of oxygen in the interior. In the stabilized polymer the oxidation rate is much slower than that in the unstabilized PP so that oxygen can diffuse into the interior without being consumed, replenishing that lost by reaction and allowing reactions requiring oxygen to continue. The absence of a high molecular weight tail means that there is no evidence for large scale crosslinking in the stabilized polymer, which is consistent with the suggestion made above that crosslinking is favored when the concentration of oxygen is very low. Crosslinks occur when long chain radicals react together and the likelihood of this happening will inevitably be enhanced if there is no oxygen available for competing reactions and reduced in the presence of a radical scavenger. For the X-EPF 30U polypropylene the addition ofTi0 2 pigment caused a large reduction in chain scission at all depths (Figure 7). After 16 weeks the molecular weight average had dropped only about 10% in the surface zone (0-0.1 mIn) and by much less at other depths. In the presence of pigment the extra addition of stabilizer did not make much difference in the measurements lllade here. It is deduced that the pigment has limited the penetration ofUV radiation and that this has been responsible for the reduction in chain scission. On comparing salnples exposed for 32 weeks with on/off cycles with a 12 hour period with those exposed for 16 weeks continuously it is noted that in the interior the fall in 11101ecu-

268

Weathering of Plastics

lar weight was greater in the cyclically exposed samples (Figure 8). Diffusion during the dark periods will have replenished the oxygen levels in the interior and the observation of lower average molecular weights is to be expected if a higher oxygen concentration favors chain scission, as suggested above. This does not explain why the molecular weight fell to a lower level near the surface when the radiation was uninterrupted than when it was cyclic. Small radicals and hydroperoxides are formed during polymer photodegradation and play an important role in the chain reactions involved. 12 They are produced predominantly near to the surface under the exposure conditions applied here, and under continuous exposure they will sustain the reaction in this locality. During periods ofdarkness it is possible that they migrate away from the surface so that when the UV radiation is turned on again the reaction will not be as rapid as it was immediately before the radiation was switched off. Conversely those reagents that migrate into deeper zones within the salnples may promote greater reaction there than happens when continuous exposure is applied, providing another reason for observing lower Mw values in the interior ofcyclically exposed samples, additional to the oxygen concentration explanation given above.

CONCLUSIONS Oxygen starvation limits molecular degradation in the interior of unstabilized polypropylene except in thin sections (say <1 mm) when tropical UV exposure levels are used. Under these conditions the reactions near to the surface are so rapid that oxygen is consumed before it can diffuse far into the material. Degradation under low concentrations of oxygen seems to favor crosslinking rather than scission and an average molecular weight measurement may mask the extent of change when both scission and crosslinking are present. The presence of stabilizer reduces molecular degradation near to the surface but reduced oxygen consumption means that more oxygen reaches the interior and reaction there is greater than in unstabilized PP. The oxygen levels are then apparently high enough for scission to dominate over crosslinking. Addition ofTi0 2 pigment causes a strong reduction in molecular degradation, presumably because it blocks UV penetration and limits photodegradation to a narrow zone at the surface. Cyclic exposure allows oxygen levels to recover during dark periods and probably allows the redistribution of other reactants (which are the products of other photodegradation reactions). This causes the degradation to be enhanced in the interior and slightly less near to the surface.

ACKNOWLEDGMENTS TJT acknowledges EPSRC for providing a studentship. The authors are grateful to Montell UK for providing the materials used in this study.

Photo-Degradation Depth Profile

269

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

L Audouin, V Langlois, J Verdu & J C M de Bruijn, J. Matel: Sci., 29 (1994) 569. G C FUTIleaUX, K J Ledbury & A Davis, Polym. Degrad. Stab., 3 (1980-1) 431. B O'Donnell, J R White & S R Holding, J. Appl. PO~1'm. Sci., 52 (1994) 1607. B O'Donnell & J R White, Po(vm. Degrad. Stab., 44 (1994) 211. Li Tong & J R White, Polym. Degrad. Stab., 53 (1996) 381. B O'Donnell & J R White, PoZvm. Degrad. Stab., 44 (1994) 211. B O'Donnell & J R White, J. Matel: Sci., 29 (1994) 3955. P Brennan & C Fedor, 43rd Ann. ConfComposites Inst., SPI, Session 23-A (1988) p.1. A V Shyichuk & J R White, unpublished. A Davis & D Sims, Weathering of Polymers, Appl. Sci. PubIs., London, (1983). J R White & A Turnbull, J. Matel: Sci., 29 (1994) 584. J R White & N Ya Rapoport, Trends in Polymer Science, 2 (1994)197.

New Light Stabilizer For Coextruded Polycarbonate Sheet

James H. Botkin Ciba Specialty Che111icals Corporation, Additives Division, Tarl)Jtown, NY, USA Andre Schmitter Ciba Specialty Che111icals Inc., Additives Division, Basel, Slvitzerland

INTRODUCTION Polycarbonate sheet has found use in a variety of architectural applications requiring long-term weatherability, including greenhouses, skylights, transit shelters, and covered walkways. Twin-wall sheet is becoming especially popular due to its exceptional strength-to-weight performance and its design characteristics. 1 The traditional approach to producing a weatherable polycarbonate sheet is by bulk-stabilization with a benzotriazole UV absorber (e.g. BZT-1, BZT -2) at a loading level of 0.2-0.5%. However, sheet stabilized in this manner shows significant yellowing after as little as 12-18 months of outdoor exposure. A n1uch more weatherable sheet can be lnade by concentrating the UV absorber near the exposed surface. For example, this can be done by coextrusion of a thin layer (typically 20-50 Jlm) of polycarbonate containing a high loading (5-10%) of UV absorber over a n1inimally-stabilized bulk layer. The highly stabilized cap layer protects the bulk layer by absorbing essentially all of the incident UV radiation. The use of a nonvolatile UV absorber in the cap layer is critical, and high molecular weight products such as BZT-l are preferred. 2 More volatile lower molecular weight products (e.g. BZT-2), while satisfactory for use in conventional bulk-stabilized sheet, have the 'tendency to volatilize when used in the coextrusion cap layer. This leads to problems such as plate-out which affect the optical qualities of the sheet. While coextrusion provides a sheet product having superior weatherability vs. conventional bulk-stabilized sheet, trends in the industry are demanding longer term weatherability. Weatherable sheet products commonly carry up to a 10 year warranty against loss of light

272

Weathering of Plastics

transn1ission and yellowing. There is a need for a UV absorber providing better long term weatherability than the state-of-the-art BZT-1. The critical attributes for a UV absorber providing superior weatherability include stronger UV absorbance (especially at wavelengths where polycarbonate is most sensitive) and improved photostability. In the present study a new hydroxyphenyltriazine UV absorber (HPT-1) is introduced which meets these requirements and provides superior long-term weatherability to coextruded twin-wall polycarbonate sheet. The product also exhibits the requisite low-volatility and has only a minimal effect on melt viscosity.

RESULTS AND DISCUSSION UV ABSORBANCE In order for the cap layer to adequately protect the base layer, the UV absorber must !. -- ~~:~].: strongly absorb the radiation at wavelengths ~_~ that produce degradation. The wavelength BZT·1 . J,. sensitivity of polycarbonate towards UVj induced discoloration has been extensively < studied and UV radiation at wavelengths less than 300 run is particularly prone to produce discoloration. Mullen and Searle3 mea250 275 300 325 350 375 400 425 sured the activation spectra for solution cast l (om) and extruded polycarbonate films and found maximum sensitivity to UV radiation at Figure 1. UV absorbance spectra (20 mg/L, EtOAc). wavelengths of 280-290 run. Andrady, Fueki, and Torikai 4 also investigated the spectral sensitivity of polycarbonate to yellowing and obtained results in good agreement with those of Mullen and Searle. The UV absorbance ofHPT-1 and BZT-I in ethyl acetate solution (20 mg/L, 1 cm path length) is given in Figure 1. HPT-1 provides stronger UV absorbance at the wavelengths below 300 nm where polycarbonate is most sensitive.

0.51

oL..

PHOTOSTABILITY UV absorbers are more photostable than most organic materials but are not permanent. 5 The photostability ofUV absorbers is particularly important in applications where a long service life is required. For example, the photostability ofUV absorbers has been shown to be critical to the lifetimes of automotive coatings. 6 UV absorber photostability is also an important factor in determining the lifetime of coextruded polycarbonate sheet. As the UV absorber

273

New Light Stabilizer

60 50~'

i ~ 40 i l1i 30 -~ I 20 -l

.._----- -

~- BZT·l -e- HPT·l

8 6 YI

llt-~+=...a-...-rl9==!:!:=-,----_

t "t'

-13-3.5% HPT.l --6- 7.0o/~

HI)'I'-I'

•

10 -I

!+- CONTROL

lOr

t",-'--"- . r

SO 100 days exposure (QUV. A·Bulb)

1'*'3.5'7£ nZT·l

1~7.0% nz·r~!.

o ~,._~_--l-___"';-_--' o 100 200 300 400 500

150

kLy exposure Figure 2. Loss of UV absorbers during UV exposure in PMMA blends.

Figure 3. Florida exposure results.

Table 1. Color change (yellowness index) photodegrades, after accelerated and natural weathering Cap layer formulation No light stabilizer

3.5% BZT-I 7.0% BZT-I 3.5% HPT-J 7.0% HPT-l

t.YI, 15,000 h Xenon-arc 22.7 7.2 5.9 5.8 5.3

t.YI, 480 kLy Florida 15.8 4.2 2.5 1.8 0.9

UV transmIssIOn through the cap layer increases, allowing degradation of the base layer. HPT-1 has been reported to have superior photostability vs. other commercially-available UV absorbers, including other hydroxyphenyltriazines. 7 The loss of the UV absorbers HPT-1 and BZT-1 in an unreactive PMMA matrix during UV exposure is shown in Figure 2. HPT-1 exhibits su-

perior resistance to loss relative to BZT-1.

PERFORMANCE IN COEXTRUDED TWIN-WALL SHEET The performance of HPT-I vs. BZT-1 was investigated in coextruded twin-wall sheet (1 em thick). The UV absorbers were incorporated in the cap layer (40 /-lm thick) at two different concentrations (3.5 and 7.0%). Performance was assessed by exposing sheet samples in Florida and in a xenon arc weatherometer with a water spray (ASTM G26, Test Method A) and measuring color (as yellowness index) periodically. Florida exposure results are summarized in Figure 3 and Table 1. After 480 kLy exposure (ca. 3 1/2 years) sheet stabilized with HPT-1 in the cap layer exhibited less yellowing than sheet stabilized with BZT-l in the cap layer. Resistance to discoloration with 3.5% HPT-l in the cap layer was superior to BZT-1 at 7.0%.

Weathering of Plastics

274 15

~CONTROL

12 .

-e- 3.5"10 HPT·! -tr 7.0% 1JIl'f·}

, ;

9··

6 *"3.5% HZT·!I ___ 7.0% BZT.l!·

Table 2. Time to onset of microcrack formation during accelerated weathering Cap layer formulation No light stabilizer 3.5% BZT-l 7.0% BZT-I

-+-1- - - - - 1

3.5% HPT-l

4,000 8.000 12,000 16,000 7.0% HPT-l hr exposure (ASTM G26, method A)

Figure 4. Accelerated weathering results.

Hrs to microcrack formation 1,500 2,500 6,000 12,500 >15,000

Exposure: Xenon-arc Weather-Ometer, ASTM G26, Test Method A (with water spray).

Similar results were obtained during accelerated weathering (Figure 4, Table 1). After 15,000 hours of exposure (equivalent to ca. 10 years in Florida), sheet stabilized with 3.5% HPT-1 in the cap layer exhibited comparable yellowing to sheet stabilized with 7.0% BZT-l in the cap layer. Sheet samples were also regularly examined for the formation of surface microcracks during accelerated weathering using an optical microscope. Microcracks act as stress concentrators, thus formation of microcracks can be accompanied by a loss of impact strength. The time to onset of microcrack formation is summarized in Table 2. Sheet stabilized with HPT-l in the cap layer showed superior resistance to microcrack formation. The results of Florida and accelerated weathering show that as expected from its higher UV absorbance and photostability, HPT-1 clearly outperforms BZT-1. Sheet stabilized with 3.5 % HPT-1 exhibits resistance to yellowing comparable to or better than sheet stabilized with 7.0% BZT-l.

VOLATILITY A comparison of the volatility ofHPT-1, BZT- 1, and BZT-2 by thennogravimetric analysis (lO°C/min heating rate in nitrogen) is given in Figure 5. HPT-1 exhibits lower volatility than BZT-2 but does not quite match the performance ofBZT-1. However, since a substantial reduction of UV absorber concentration can be achieved by using HPT-I, volatility of the UV absorber is somewhat less critical.

EFFECT ON MELT VISCOSITY Incorporation ofUV absorbers at the high levels used in the cap layer produces a decrease in melt viscosity of the polymer by a plasticizing effect. The effect ofHPT-1 relative to BZT-1

275

New Light Stabilizer ,..., 2000 ...

. -+7.0% nZT·l, 270C -8-4.5% UPT-I, 270C

'" ~

80·· .

] .S 60 !!

1600

::' 1200·· .;;;

i -fi-4.5% IIPT-I,.

Cl:: 40 . ~

1-111'1'.1 BZT-l '-nZT·2

o . ' ---'-------'--.-.........- l 200 225 250 275 300 325 350 375 400 425 T (OC) Figure 5. Volatility ofUV absorbers by TGA (I O°C/min, N2).

.;;:

800···

265 C

. ;

.-_.~--+-------......;

100

1000

10000

Apparent shear rate (sec") Figure 6. Melt viscosity results.

1>2'1'·2:

on polycarbonate melt viscosity was determined by measuring apparent viscosity of two commercial cap-grade resins over a range of shear rates with a capillary rheometer at 270°C. Results are summarized in Figure 6. HPT-I at 4.5 % showed a slightly greater plasticizing effect than BZT-I at 7.0%. Thus the use of a slightly lower melt temperature may be advisable when HPT-I is used in the cap layer. When the melt temperature was reduced to 265°C, the cap resin containing 4.5 % HPT-I gave a viscosity vs. shear-rate profile comparable to cap resin containing 7.0% BZT-I at 270°C.

H1>T·l:

CONCLUSIONS Compared to BZT-I, the new UV absorber HPT-I exhibits stronger absorbance at wavelengths where polycarbonate is most sensitive and improved photostability. As a result, HPT-I provides superior weatherability to twin-wall coextruded polycarbonate sheet. Sheet stabilized with

276

Weathering of Plastics

3.5% HPT-1 in the cap layer exhibits weatherability comparable to or better than sheet stabiIized with 7.0% BZT-1. HPT-1 also features low-volatility and has only a minimal effect on melt viscosity.

ACKNOWLEDGMENTS The authors would like to thank Mr. Guy Jordy, Ms. Emerald Collins, and the Additives Analytical Research Department for their excellent laboratory work in support ofthis project, and to Ciba Specialty Chemicals Corporation for pem1ission to publish this paper.

REFERENCES 1 2 3 4 5 6 7

Press Release PR #41-98, GE Structured Products, September 4, 1998. H. Hahnsen, W. Nising, T. Scholl, H.-J. Buysch, and U. Grigo (Bayer AG), U.S. Patent 5,108,835; 1992. P. A. Mullen and N. Z. Searle, J. Appl. Polym. Sci., 1970, 14, 765-776. A. L. Andrady, K. Fueki, and A. Torikai,1. Appl. Polym. Sci., 1991, 42, 2105-2107. R. C. Hirt, N. Z. Searle, and R. G. Schlnitt, SPE Trans., 1961,1,26-30. D. R. Bauer, 1. Coatings Tech., 1997, 69,85-95. 1. E. Pickett, "Pemlanence ofUV Absorbers in Plastics and Coatings", presented at 7th Annual ESD Advanced Coatings Technology Conference and Exposition, Detroit, MI, September 1998.

Ultraviolet Light Resistance of Vinyl Miniblinds Part 2. Reaction Products Formed by Lead in Air

Richard F. Grossman Halstab

BACKGROUND Exposure of vinyl compositions to sunlight or to laboratory sources of ultraviolet light normally does not result in exudation of the heat stabilizer, whether based on lead, tin, or other metals. In particular, the resistance of lead stabilizers to nligration is well known. 1 Previously, samples of a typical rigid profile extrusion compound were exposed to UV-A and UV-B irradiation for 1500 hours in a Q-Panel QUV accelerated weathering apparatus. 2 Some of these samples were lead stabilized; others contained a tin mercaptide stabilizer. In no case did surface lead increase above the error in detection by the atonlic absorption procedure used, 0.01 ~lg/cm2. On the other hand, the same exposure of lead stabilized vinyl miniblinds led to detectable quantities of surface lead, leveling offat 0.1-0.2 ~g/cn12. This difference in behavior may be a reflection of the very high filler loadings (as lTIuch as 80 phI' CaC0 3 ) used by sonle miniblind manufacturers. The lead compound that exudes to the surface appears to be the reaction product of the stabilizer, tribasic lead sulfate, with HCI, that is, mono- or dichlorotribasic lead sulfate. Lead stabilizers and their HCI reaction products are highly insoluble in dilute hydrochloric acid. 2 Concentrated nitric acid was required to extract the lead from the miniblind surface after UV light exposure. It is interesting to note that in their study, the Consumer Safety Protection Agency (CPSC) found complete equivalence using conc. HN0 3 or dilute aqueous HC1. 3 The dilute HCI, at 37°C for 6 hours, was intended to mimic human digestion. To permit samples to be run more rapidly, conc. HN0 3 was substituted. This yielded the same results. What they dissolved for analysis was, therefore, primarily not insoluble lead stabilizer or its similarly insoluble reaction products, but some readily HCI-soluble lead compound. There are such compounds widely available. Basic lead carbonate, the comnlon constituent of airborne lead-containing dust,4 is quite soluble in dilute HCI (as well as in conc. HN0 3).

278

Weathering of Plastics

Many vinyl miniblinds in the field have been found with surface lead exceeding 0.1-0.2 Jlg/cm2.5 Levels of 1-10 Jlg/cn12 are common, with some specimens having levels as high as 40-80. Indeed, CPSC found surface lead levels well above 0.1-0.2 flg/cm2 after exposure of clean n1iniblind surfaces to UV light (using, of course, conc. HN0 3 to dissolve the product). Their technique, however, involved exposure for a length of time, cleaning the surface, and re-exposure, with their results summing the lead found after each successive step. There is no way that CPSC could have foreseen that such treatment would remove the protective surface layer that the previous exposure generated. Nor is it likely they would have been aware of the insolubility of lead stabilizers in dilute hydrochloric acid.

EXPERIMENTAL An unhighlighted but intriguing feature of the CPSC data was that one ofthe highest levels of miniblind surface lead found in the field (in North Carolina) came from a site where the vinyl contained no detectable lead. 2 The obvious conclusion, that the miniblind was not lead stabilized, was rejected by North Carolina Dept of Health & Natural Resources (NC-DEHNR) in favor of: (A) all the lead must have exuded to the surface, therefore none was left in the miniblind; or (B), maybe it was a different miniblind than the one whose surface was analyzed. 5 It was, therefore, with some relief that a genuinely tin-stabilized miniblind was found (in NW Indiana) that apparently had not been cleaned in 2-3 years. Slats from this miniblind contained no detectable lead (by AA) but about 0.2 wt % tin. (It is thus likely that they were manufactured in North America). The surface, however, had 40-50 Jlg/cm2 lead. SEM-XRF indicated that most of the surface coating was calcium carbonate. Interspersed were large rhombic crystals that appeared to grow from the vinyl surface. These were large enough (5-10 11m) for detailed analysis and proved to contain Pb, Ca, and CI in the ratio of 1: 1:3, and did not correspond to any previously known compound. 6 A test compound was prepared, corresponding to rigid conduit: PVC 100, CaC0 3 40, impact modifier 3, processing aid 2, ester lubricant 2, stabilized with 2 phr of 65/35 di- to monobutyltin isooctyl thioglycolate. Strips of 1.5 mm thickness were exposed to UV-A radiation at 50°C in a continuous moving stream of air (circa 5 l/min) that was previously passed over finely divided (2-5 11m) basic lead carbonate. Under these conditions, in 300 hours of exposure, 50-65 I1g/cm2 of surface lead developed. This coating proved soluble both in dilute hydrochloric and conc. nitric acids. It again contained Pb, Ca, and CI in the ratio of 1: 1:3. Equimolar quantities ofCaC0 3 and basic lead carbonate were dissolved in hot IN HCl. Slow cooling yielded large crystals (1-3 lllin) of the above compound (containing Pb, Ca, and CI in the ratio of 1: 1:3).

279

Ultraviolet Light Resistance

Similar experiments were carried out using the same compound as above, but without CaC0 3 filler. One phr Ti0 2 was used instead to provide opacity. After 300 hours ofUV-A exposure, 50-65 f.1g/cm2 of surface lead also developed (again despite the absence of a lead stabilizer). This product was also soluble in dilute hydrochloric acid, contained Pb and CI in a 1: 1 ratio, and appeared identical to reference samples of basic lead chloride.

DISCUSSION The most common naturally occurring fonn of basic lead carbonate, hydrocerussite, [2PbC0 3 .Pb(OH)2], corresponds to:

II O-C-O- Pb-OH Pb/

"O-C-O- Pb-OH II

o Reaction with HCI generates CO 2 and basic lead chloride [Pb(OH)CI]. The latter is reasonably light stable, at least in comparison to PbCI 2, which rapidly loses Cl 2 to leave colloidal lead, much like AgCl. Thus basic lead chloride as a stable end product from the settling, or static attraction of lead dust, in air is not surprising. It is well known that PbCI2, although not a strong Lewis acid, forms double compounds with alkali metal and alkaline ealth halides, e.g., CaCI2.PbCI2.7 This is apparently also the case with basic lead chloride. For example, we find a 1: 1 addition compound with hexa-chloro-l,3-butadiene [C 4CI 6 .Pb(OH)CI] that appears highly resistant to UV light degradation at 50°C. The reaction product in the presence of CaC0 3 appears to be the double compound CaCI2.Pb(OH)Cl. This is quite interesting since CaCl 2 does not otherwise appear, despite the prevalence of CaC0 3 at the degraded vinyl surface. Certainly one must suspect that the presence of the lead salt is involved, and thus consider generally the extent to which stabilizers may be able to transfer chloride to receptive "filler".

CONCLUSIONS It is likely that most, if not all, of the lead-containing detritus found on the surfaces of vinyl miniblinds results from the conversion of lead dust in air to chlorinated products, principally basic lead chloride, from the HCI produced by UV-light assisted degradation of vinyl. These

280

Weathering of Plastics

conversion reactions probably serve to retain lead-containing compounds accumulated from the air, as compared to chemically neutral surfaces (e.g., painted or coated wood or aluminum equivalents). There seems to be no need for the presence of lead stabilizer in the compound for the surface lead accumulation reactions to occur. Although there is no requirement in terms of desired properties to use lead stabilizers in vinyl miniblinds, the widely publicized conclusion that such stabilization adds to the hazard of heavy metal exposure is simply not justified by experiment, and probably resulted from consideration ofdata from the field using too narrow a technical base.

REFERENCES 1. 2. 3. 4. 5. 6. 7.

The Environmental Itnpact of Lead Stabilizers, Nordic Plastic Pipe Association, Stockholm, Jan. 1995. R.F. Grossman and D. Krausnick, JVAT, in press. W.K. Porter, CPSC Report, Miniblind Lead Investigation, Sept. 18, 1996. Ter Haar and Bayard, Nature, 232, 553 (1971). Private comluunication, North Carolina Dept. of Environment, Health and Natural Resources. J.\V. Mellor, Inorganic and Theoretical Chemistry, Vol. VII, Longmans, Green & Co. D. Greninger et al., Lead Chemicals, ILZRO, New York, 1975

Case Studies of Inadvertent Interactions Between Polymers and Devices in Field Applications

Joseph H. Groeger, Jeffrey D. Nicoll, Joyce M. Riley, Peter T. Wronski Altran Materials Engineering, a Division ofAltran Corporation, Canlbridge, MA, USA

INTRODUCTION Polymeric compounds are selected for a wide range of applications by technical persons with a variety of backgrounds. Initial choices may be moderated by other specialists who are often unaware of the potential pitfalls and adverse interactions associated with the use of cost-effective or inappropriate alternate materials. Manufacturers who provide subcomponents may not be included in the design reviews of finished products into which their components are being used. Additionally, suppliers of commercial polymeric n1aterials may be unaware of how their materials are being applied. As a result of these and other considerations, materials selections may be made based on a review limited to basic engineering properties. Considerations oflong-tenn perfoffi1ance and response to specific operating conditions requires a degree of attention and insight that may be overlooked. Several case histories are cited in which some aspect of materials selection and design were deficient in the application. A thennally activated electrical switch fonnerly made with a phenol formaldehyde thermoset resin was redesigned to include a thennoplastic resin. Localized heat associated with the arcing activity of the switch contacts caused thermal erosion of the housing, releasing reactive sulfur compounds which then reacted with the electrical contact faces, causing irregular performance and eventual contact welding. A pressure relief device in a consumer product was found to have highly variable performance as a result ofextensive processing aid additions to the base polymer, selection of poor quality raw materials, and no attention to a root cause analysis with a review of the compound. Plasticizers released from PVC wire insulation at elevated operating temperatures wicked along the conductor strands and onto relay contacts, resulting in a power plant shutdown. Components from a pharmaceutical product container were found to be exuding phthalate compounds which were not expected based on an initial review of the raw materials.

282

Weathering of Plastics

These cases are presented as constructive examples for those seeking to maximize the performance and useful life of devices making use of polymeric components through an integrated materials selection and design approach.

DECOMPOSITION OF THERMAL SWITCH A thermal limit switch used in a number of domestic and commercial appliances was historically manufactured with either a ceramic or a crosslinked phenol formaldehyde, providing many years of reliable service. A change in materials had been implemented to facilitate processing, resulting in a housing made with thermoplastic polyphenylene sulfide (PPS). The housing contained silver-laminated bronze electrical contacts, one of which was mounted on a bimetallic arm to provide thermally-controlled switching action. Failures of this switch were encountered wherein the contacts were found to weld together, resulting in a thermal runaway condition caused by a failure to interrupt current to the heater that the switch was intended to control. A foren'. sic review of representative failed switches was undertaken. Figure I presents a scanning electron micrograph of the surface of a contact reFigure I. Surface of contact showing raised areas where moved from a failed switch. On the surface, welding occurred, 150x. many melted areas are clearly visible. Some ofthese are flat, showing the previously molten condition of the metal contacts. Metallographic crosssections through such a contact showed severe localized melting. Elemental analysis of the contact surface indicated that silver sulfide was present. This compound produced an insulating layer on the surface of the contact, resulting in erratic current flow and localized heating due to limitation of the available contact surface area. Switches in various stages of degradation were operated with thermocouples placed on the contacts and housing. Measurements indicated significant resistive heating, merely due to flow of the rated current. Chemical analysis ofthe polymer heated to the as-found level, using gas chromatography and mass spectrometry (GC/MS), confirmed formation of hydrogen sulfide, carbonyl sulfide, sulfur dioxide, hydrogen, and methane. Examination of the switch housing interior surfaces surrounding the contacts revealed significant erosion of the polymer as shown by the light colored oval region in Figure 2. Closer examination revealed the glass and mineral reinforcement particles within the PPS

Case Studies

283 compound standing in relief, due to polymer pyrolysis. This damage was due to the intense localized heating produced by arcing as electrical contact between the switch contacts was established then broken during normal operation. The combined evidence of contact melting and PPS pyrolysis suggested short-term temperatures in excess of 600°C.

INCONSISTENT PRESSURE RELIEF MEMBRANE A pressure relief membrane used in a consumer product was found to exhibit erratic performance both in quality assurance testing and in the consumer market. The pressure relief device was a critical component and played an integral role in product function and safety. The device was manufactured using a compounded thennoplastic polypropylene which was injection molded into the necessary form. As can be seen in Figure 3 the molded part is quite complex in design; consisting of numerous ribs, radial formations and most importantly, the thin membrane which acts as a pressure rupture diaphragm. Figure 3. Top view of pressure relief device. The latter is coined in the injection molding process. Investigation of the device revealed many areas of misapplied designs and a general focus on processing performance instead of functionality. The thermoplastic compound which was used to fabricate the units made use of a fairly complicated fonnulation. The original base resin was dropped from the supplier's product line and alternates were substituted. In conjunction with these changes, increased device anomalies and difficulties controlling the burst pressure range were experienced. After a preliminary materials investigation ofthe disclosed formulation, interactions of the materials being used were identified as being inordinately complex and in some cases inappropriate for this application. Figure 2. Interior surface of switch housing showing polymer, erosion, 9x.

284

Weathering of Plastics

Organic chemical analyses of representative devices were conducted using GC/MS. This method was selected to confirm the identity of the organic ingredients and processing aids in the questionable formulation. GC/MS analysis of the seals revealed significant formulation variations between different lots ofmaterial. It was determined that the use of additives such as the antioxidants, antiblocking agents, internal lubricants, and other processing aids was inconsistent. The most significant variations were among materials not specified in the formulation. Processing aids such as silicones (used as internal lubricants to modify flow behavior), plasticizers (typically used for increasing impact resistance and adding flexibility), and waxes (used as lubricants and flow modifiers) were noted to be present in many of the device lots. These components appeared at random and were not used consistently. It was suspected that they were added as on-line processing aids to assist with mixing by the compounding operators and/or to achieve a target n1elt flow index. The formulation suffered from years of incremental modification for performance and processing issues which often suppressed the symptoms but never addressed the root causes. For exan1ple, there were three agents listed in the formulation which served as antioxidants. Due to the nature of their chemical functionality, these materials did not enjoy a positive synergy. Instead they competed in the formulation causing none of these materials to offer as much protection to the resin and other organic components in corrlbination as they would when used individually. The antioxidant package was further complicated when a review of their functional characteristics was completed. Originally, the molded pressure relief device suffered from a reaction with copper within the contacting unit surfaces. A metal deactivating antioxidant was added to the formulation to correct this problem. A review ofthe formulation clearly indicated that the original antioxidant was an amine (nitrogen-hydrogen) compound. This antioxidant sustained limited thermal decomposition during processing, leading to the production of amine compounds. These reacted with copper, leading to the formation of blue-colored copper compounds. While the addition of the metal-deactivator was successful in reducing this occurrence, the original antioxidant was left in place. The replacement and original antioxidants were not chemically compatible, nor was the amine antioxidant stable with respect to the antioxidant supplied in the base polypropylene resin. A third antioxidant was then added to improve oxidative stability. A different problem was noted when a scanning electron n1icroscope (SEM) was used to examine selected areas ofrepresentative seals. The high magnification ofthe SEM provided a view of the relative size ofthe individual filler particles and their alignment in key areas such as the diaphragm. Examination revealed that the filler materials had a tendency to agglon1erate in this region and that the overall filler concentration in the diaphragm area was inconsistent throughout many devices. As shown in Figure 4, the talc particles were quite large when compared with the overall thickness ofthe diaphragm. As illustrated in this micrograph, the particles aligned in the plane of the n1errlbrane and created a stacking effect. In this

285

Case Studies

Figure 4. Micrograph of diaphragm cross-section, 605x.

Figure 5. Micrograph of diaphragm comer, 226x.

case, the shape of the particles was inappropriate due to the flow mechanics in the mold cavity. Figure 5 shows the comer at the edge of a representative diaphragm. The filler particles in this area were also dramatically aligned along the curvature of the diaphragm. This suggested that the resin flow in this area during molding was restricted by the presence of the talc particles. This caused the residual stresses in the diaphragm area to be quite high and the particle size of the talc to vary depending on the level of flow restriction during injection. The effect of the talc particle size variation on the inconsistent performance of the seals was significant. This characteristic mainly affected the flow rheology ofthe compound under high shear conditions during injection molding. The talc particle size, in comparison to the diaphragm thickness, also lead to an erratic influence on the tear characteristics during product performance. Talc agglomeration and absence of bonding with the base polymer further contributed to poor performance. Inconsistent diaphragm burst performance was caused by a combination of chemical, physical and rheological phenomena. The lots of devices which exhibited a particularly high burst pressure were the result of a very fine particle size talc in conjunction with a low concentration of processing aids. The increased strength of the base resin and lack of large talc particles for burst initiation necessitated high burst pressures. The devices which exhibited lower diaphragm burst pressures suffered from a combination of large talc particles and an absence of lower molecular weight polymer to assist with the flow and wetting of the filler. This resulted in high orientation effects which led to very high residual stresses causing premature failure. These anomalies illustrate the combined effects of the uncontrolled chemical additives and random talc particle size on the consistent performance ofthe compound. In this formulation, even if extreme care were taken in manufacturing, the number of materials involved and the inherent variability and performance of the talc made it virtually impossible to produce a consistent product.

286

Weathering of Plastics

PLASTICIZER BLOOM FROM PVC CABLE JACKETS During inspections at a nuclear power plant, green liquid deposits were found concentrated on the surface of selected low voltage cables, at their terminations as well as in the instrument panel in which these cables ended at connections. The cables were rated at 600 volts and incorporated a cross-linked polyethylene (XLPE) insulation with a polyvinylchloride (PVC) jacket. The estimated age of the cables was 20 years. The green liquid deposits were determined to be non-drying, with a high viscosity, and good lubricity. Analysis of this liquid by Fourier Transform Infra-Red Spectroscopy (FTIR) confirmed that it was mostly adipic acid diethyl ester. This compound is a common plasticizer for PVC and is typically yellowish in color. An FTIR absorption peak unaccounted for by adipic acid diethyl ester was assigned to a silicone fluid (diphenylsilane). This may be attributed to a second plasticizer used in these cable jackets. Samples of the liquid were pyrolized and the residue was analyzed with energy dispersive X-Ray analysis (EDX). This revealed the presence of copper with traces of aluminum, silicon, calcium, iron, and lead. The presence ofcopper salts in the fluid was responsible for the noted green color. The presence of these green fluid deposits closely followed a record 'heat wave' in this particular region. It was deduced that this elevated regional temperature caused the sudden appearance of these exuding plasticizer compounds from the PVC cable jackets. These compounds can cause severe consequences in electrical systelns due to their insulating properties. If these compounds were allowed to nligrate into electrical switches, relays, or meters they would inhibit proper performance. In this particular case, the plasticizer impinged on the jackets of adjacent cables, causing them to swell then split. In another identical occurrence, a plant shutdown resulted when plasticizer crept onto the surface of electrical contacts used for a punlp motor relay.

EXTRACTS FOUND IN PHARMACEUTICALS The presence oftwo plasticizers, dioctyl phthalate (DOP) and diisooctyl phthalate (DIOP), in a drug formulation caused significant concern to the pharmaceutical companies since aromatics of this type are under regulatory scrutiny. Investigation into the origins of these contaminants led to analytical review of elastomeric components of the product container. Extensive GUMS analysis isolated the source of the DIOP as being the elastomer raw material. Further research indicated that the supplier of this elastomer was adding DIOP during manufacture to act as a melt-flow modifier to control the Mooney Index of the final product. The DOP, however, was traced to contamination from the polymer compounding equipment. Frequently, oils used to lubricate mixing equipment exude into the compound being produced through dust seals, for example. Knowing this, manufacturers will often utilize lu-

Case Studies

287

brication products which are compatible with the polymer products that are produced. In this case, however, the oil utilized contained DOP which would be acceptable in many thennoplastic compounding applications, not slated for medicinal use. The resultant extraction of DOP from components of the product container, however, was not acceptable.

DISCUSSION Development of thermoplastic and thermoset polymer cOlnpounds is a mature science that continues to grow with the developlnent of new types of additives, changing regulatory requirements, and proprietary considerations. The selection of all materials that are incorporated into a cOlnpound may follow lines that are not always clear. Some ingredients may be outdated. Others may have been added for a customer-specific end use and the compound later became available for the general market. A very wide range of off-the-shelf compounds are available for engineering applications. Many will fit into the existing requirements or designs and/or processes may be altered to accomnl0date the compound that best fits the needs. These choices, though, are often limited to the general engineering/technical properties without sufficient detailed consideration of the materials in context of the application. An ideal situation is one in which the end-use nlanufacturer has available the equipment necessary to develop a polymer compound specifically suited to an individual application. In this clean sheet approach, each ingredient may be carefully considered in context ofthe application, aging characteristics, processing effects, and synergy with other formulation conlponents. Conlpounding facilities need not be directly available; contract organizations are available and many of the commercial polymer compound suppliers offer custom compounding services. Analyses of plastics failures and contamination issues often indicate that it is necessary to return to the basics and re-examine the material in context of the application. With this approach, a polymer would be formulated using a minimum number of ingredients, each of which would be the most appropriate and efficient for the end use. By reducing the nUluber of ingredients, the controls necessary for each supplier are greatly silnplified and the potential for adverse interactions reduced. Many raw materials are more complex than may be apparent and, in some cases, the 'hidden' ingredients may be detrimental to an application. Virtually all commercial elastomers are supplied with an antioxidant already included and the type may change periodically. Masterbatching agents and processing aids, such as calcium stearate, may be used when adding antioxidants to a raw polymer. Crosslinking additives and their synergists are another source of antioxidants and other cOlnpounds. Crosslinking is a chemically challenging process in which thermal decomposition of a reactive peroxide is typically used to provide free radicals. This requires an additional antioxidant to protect the polymer,

288

Weathering of Plastics

while reaction products including acetophenone, cumyl alcohol, and acetic acid become available to interact with the other raw materials or additives. In the case of the thermal limit switch, the choice of materials for the housing inherently led to a reduction in the useful life of the device. The stability and useful life of the switch could be readily enhanced through the use of a polymeric housing that does not produce reactive gaseous products. Many thermoset materials are available, as are ceramics. While the near-term economy of using a thermoplastic material may have appeared attractive, the long-term effect on performance may not have been readily apparent when a material substitution was made. In the second situation corrective measures were implemented so that predictable and consistent performance of the pressure relief device could be attained. Compound reformulation took place which included the careful selection of a clean hom*opolymer base resin, a specially designed and compounded antioxidant and a low aspect ratio, small particle size reinforcement. The compound simplification, in combination with highly functional components, allowed for exceptional performance and reliability. In the example of plasticizer bloom from a set of cables, it is interesting that the simple loss of a compounding ingredient could lead to such indirect, but major consequences. In this case, exposures to long-term conditions of elevated temperature could be surmised, based on the application and service environment. Grafted plasticizers are available. Alternatively, though, a complete reconsideration of the material in this environment would have been beneficial. A polymer compound that is inherently flexible would not involve a plasticizer and the potential adverse effects of its loss. Finally for the pharmaceutical container component example, reformulation of the raw polymer compound, as well as substitution of machine lubricant with a food-grade aliphatic mineral oil was necessary, followed by substitution of increased purity raw materials, before use of this material could be continued.

CONCLUSIONS It is important that the total life cycle ofpolymeric compounds be considered in context of the end-use application. Some basic guidelines can be developed from a review of situations in which the process was not optimized. • The application should be well understood in terms of stresses (thermal, chemical, physical, radiation, etc.). Near- and long-term exposures must be considered. • Review the candidate or existing material with a fresh perspective and careful attention to all raw n1aterials, their quality, and roles.

Case Studies

289

• Simplicity of design facilitates processing, cost control longevity, and quality assurance. This requires that raw materials be inherently suited for the compound, rather than placing a strong reliance on complex additive packages. • The perfonnance of a compound cannot be limited by processing or post-processing handling. Attention to detail is necessary to assure that the intended and realized formulations are identical. • While some of these suggestions may seem tedious, it is often the case where short-tenn economy and lack of application-specific insight may lead to significant losses when a poorly selected material fails in service.

REFERENCES 1 2 3 4 5 6 7 8 9

Hoffman, Werner (1989). Rubber Technology Handbook. New York, NY. Hanser Publishers. Bhowmick, Anil and Howard Stephens, Eds (1988). Handbook of Elastomers. New York, NY. Marcel Dekkel; Inc. Schnaebel, Wolfram (1981). Polymer Degradation: Principles and Practical Applications. New York, NY Hanser Publishers. Sekutowski, Dennis (1992). "Inorganic Additives". in Plastics Additives and l\1odifiers Handbook, Jesse Edenbaum, Ed. New York, NY. Van Nostrand Reinhold. Gachter, R. and H. Muller (1993) Plastics Additives Handbook. Cincinnati, OH. Hanser Gardner Publications. Charrier, JM (1990). Polymeric Materials and Processing. New York, NY. Hanser Publishers. Barth, H. and Mays, J. (1991). Modem Methods of Polynler Characterization. New York, NY. John Wiley Publishers. Engineering Plastics and Composites (1990). Metals Park, OH, ASM International. Rauwendaal, C. (l991).l\lixing In Polymer Processing, New York, NY. Marcel Dekkel:

Automotive Clearcoats

George Wypych Che111Tec Laboratories, Inc., Toronto, Canada

Fred Lee Atlas Electric Devices Co 11 Ipa n)l, Chicago, USA

INTRODUCTION Preceding chapter indicated the need for specific infonnation required to design experiment of material weathering. The aim of this paper is twofold: • generate and systematize infonnation on degradation behavior of automotive coatIngs • provide an example of data selection in preparation for weathering studies The first reason is driven by the fact that such review of technology was not presented so far in spite of the fact that clear coats are of interest of many groups in industry, testing, and university research, including: automobile, motorcycle, bicycle, manufacturers; manufacturers of coatings for repairs; Inanufacturers of exterior metal parts; manufacturers of exterior plastic parts; manufacturers of polymer blends for auto1110tive applications; compounders of plastics; niche markets for clear coats (office furniture, shelving, lighting fixtures, tool boxes, doors); raw material suppliers for coating manufacturers (polymers, curatives, stabilizers, catalysts, initiators, rheological additives, pigments); research institutes (development ofnew products, methods of testing, raw materials used for coatings); national testing institutes; standardization organizations; commercial testing laboratories; university research (development of new products, methods of testing, raw materials for coatings); environmental institutes (studies on environmental impact of degradation products); corrosion protection (research, 111anufacturers of protective chemicals); consultants in the area of weathering and ISO 14,000; military (research, engineering, quality control); aerospace (all aspects of exterior applications of coatings and plastics); others working in the similar fields. This long list shows that the number ofpeople and institutions involved is very large thus a comprehensive review ofinfonnation that is currently scattered is required. As a long list of

292

Weathering of Plastics

references shows, the currently available information is available in many sources - some of which are difficult to obtain. The information provided in this chapter should be updated in the next two years concerning that a very broad research on powder coatings is under way which will affect provided here list of materials used and the list of potential mechanisms of degradation. For the second purpose of this chapter, it is important to mention that the choice of automotive industry is ~dequate because it size warrants a large number of quality research and thus data. This allows to review all aspects of data required prior to weathering testing. It is also important to note that automotive coatings were recently developed from prone to failure technology to robust process which yields durable products. This successful conversion occurred in spite of the fact that the process was complicated by additional needs to eliminate or limit use of solvents which imposed many restrictions on the development process. It s also impoliant to note that there are still large gaps in understanding which this contribution tries to point out to generate required research.

APPLICABLE STANDARDS EXPOSURE IN LABORATORY DEVICES

Table 1. Automotive exterior coatings • applicable standards for the laboratory testing Standard

Equipment

lrradiance, W1m2

Temperature °C

RH,%

SAE J1647

HID chanlber

38-47

SAE J1960

Xenon-arc (water)

0.55 @340

38 and 70

95 and 50

SAE J2019 SAE J2020

Xenon-arc (air)

80 @300-400

38 and 47

95 and 50

Fluorescent UV

0.43 ~310

VDA 621-4 (Gennan)

Xenon-arc

70 UV/SO dark 63 UV/I0 dark

LP-463PB-16-0 1 (Chrysler) LP-463PB-9-0 1 (Chrysler) BO 101-1 (Ford) GM9125P (GM)

Carbon-arc Humidity chamber Carbon-arc Carbon-arc Fluorescent UV

63-71 none

37.2-38.4 60-65 (BP)

98-100

60±2 70 UV/SO condo

Xenon-arc MO 135 (Nissan) BS AU 148 (British) JIS D0205 (Japanese)

Carbon-arc Xenon-arc Mercury lanlp Carbon-arc Xenon-arc

63±3 or 83±3

50 and 90

89±3 and 38±3 63±3 or 83±3 63±3 or 83±3

50±5 50±5

Automotive Clearcoats

293

OUTDOOR EXPOSURE SAE J 1976 applies to outdoor exposures of automotive coatings and other exterior materials. Coating systems are exposed in panel racks (unbacked exposure) and black boxes.

SOLAR FRESNEL REFLECTOR APPARATUS SAE J 1961 applies to the use of concentrated radiation for exposure of automotive samples including coatings. Apparatus should be operated in dry, sunny climates receiving 3000-4000 h of sunshine. In addition to exposure during the day, specimens are sprayed in the night for 3 min in each 15 min. Two types of exposure are used: non-insulated and insulated (backed with plywood). In insulated exposure, the insulation is only used between November 1 and March 31.

SUMMARY It is interesting to note that the national standards are not playing an essential role in testing of automotive coatings. Only Britain and Japan have national standards. The British standard is old (1969) and probably not frequently used. The laboratory testing is mostly based on SAE standards which allow for the use of all three weathering devices (carbon-arc, xenon-arc, and fluorescent UV). It is important to note that only Xenon-arc device offers full control of all weathering parameters (irradiance, temperature, and humidity) which are specified in the SAE standards.

GENERAL DISCUSSION OF TRENDS Quality of automotive finishes, legal requirements, and environmental concerns were the driving factors for changes in automotive coatings. 1 During 1950-1970, oven-dried alkyd-melamine, lTIOnOcoat, straight-shade, coatings were in the common use. In the period of 1970-1990, the evolution of paint technology was gradually leading toward a more complex systen1 of autolTIotive finishes which eventually included low-solids, solvent-borne basecoats and alkyd-melamine clear coats, later replaced by high-solids basecoats and acrylic-melamine clear coat with UV/HALS stabilizers. These systems included metallic basecoat. During the 1110st recent times, several new solutions were introduced, including water-borne basecoats with urethane clear coats. Even more recently, water-borne basecoats were combined with powder clear coats. The above short introduction indicates three major trends:

294

Weathering of Plastics

Period

Action

Drivers

1950-1970 development of new technology of coatings

quality

1970-1990 development of clear coat technology

quality, appearance, durability

1990-

envirOlunent, legislation

development of water-borne and powder coatings

The period of 1970-1990 was especially instructive in stressing importance of testing with a special emphasis on weathering testing. During this time, many failures occurred, indicating that both long-term perfonnance predictions and quality control must include weathering testing, considering that failure is very expensive.

PERFORMANCE CONDITIONS Automotive coatings meet variable environmental conditions due to the widespread use of cars in different climatic conditions. Table 2 gives a list of essential parameters.

Table 2. Typical parameters of performance of automotive coatings. Parameter UV radiation

Average value wavelength: 295-380 nm irradiance: 0.35 W/m 2 @340 nm

Maior influences photochemical conversions photooxidation degradation of metallic effect

Telnperature as a function of air -60 -:- 100°C (up to 115°C) telnperature, IR, and color

conlbined degradation activity increased rate of reactions caused by other parameters

Humidity

stress due to thermal movement hydrolysis

10 -:- 100%

non-oxidative photodegradation mar (acid etch) stress due to change of volume Wetness

1-40% total time

extraction hydrolysis penneation to interface

Pollutants and fog

pH of fog as low as 2

surface erosion mar (acid etch) hydrolysis crack initiation

Acid rain (dew)

pH as low as 1 pH of de\v as low as 2

surface erosion lnar (acid etch) hydrolysis crack initiation deposition of salts into clearcoat

295

Automotive Clearcoats

Parameter

Average value

Major influences

Dust particles

widely variable

absorb moisture and acids embedment into clearcoat

Salt (deicing, coastal)

surface etching delatllination corrosion shrinkage

Evaporation of volatile C0111pOnents

surface roughening crack initiation

Pancreatine (bird droppings)

surface etching

MODES OF FAILURE Table 3 gives a list and analysis of modes of failure. Table 3. Modes of failure of automotive coatings in relationship to causes and essential parameters of weathering involved in the failure. Mode of failure Causes Parameters Gloss loss photoxidative processes caused by combination of parameters; correlation UV wavelength (18 months in Florida)38 strongly depends on the control and simulation of conditions of degradation;2o,21 irradiance level initial loss is due evaporation of volatiles 22 (1700 h Xenon arc)38 temperature humidity shrinkage Yello\ving (2500 h Xenon arc i 8 Adhesion loss (2 years in Florida)21

chemical conversion of certain chemical groups in some resins; sonle hardeners UV radiation increase probability;38 more visible with lighter (white basecoat) colors temperature partially attributable to photochemical processes but becomes visible due to UV radiation stress causing delanlination (sources of stress - variable temperature and temperature tnoisture intake) moisture pH

Cracking

see adhesion, water spots, and surface erosion

see adhesion loss

(2 years in Floridai l

Mar (a few months )26

fort11ation of fine scratches due to the environtnental effects (associated defects: UV radiation defonnation and spotting); car washing, in-plant polishing and exposure are precipitation (pH) main causes; typical reasons are photochetnical damage, droplet's swelling, and abrasion solid particle deposits 19,26 H 20 concentration

Water spots

occurs due to deposition of inorganic salts into the surface of clear coats (initial acid rain (pH) UV exposure under dry and cool conditions limits the process );25 fonnation of hydrolysis microscopic blisters and clear coat cracking fonns the so-called defect of UV radiation "unrenlovable water spots,,17

temperature

296

Weathering of Plastics

Mode of failure Surface erosion

Causes Parameters acid rain in combination with dust collection (dust absorbs pollutants) and UV radiation photooxidation; pancreatine related surface damage mostly occurs \vith freshly oxygen produced cars 1 dissolved acids pancreatine

Oil staining

polluted lnotor oil containing carbonaceous products of degradation 38

used oil

Substrate con-osion

loss of barrier properties, transport of ions to interface with nletal

deicing salt salnvater particles

The above list of modes of failures indicates that failure is generally caused by a con1bination of factors which sets the important criteria of testing: • parameters of exposure must precisely imitate conditions of performance • reproduction of conditions depends on the precise control of several parameters (not just UV radiation) • method of exposed specimen testing determines result. The length of time to encounter failure is given as a general infolmation to illustrate premature failures of selected formulations.

CHEMICAL COMPOSITION Automotive coatings are applied for two groups of substrates: metal and plastic. The following diagrams best explain component elements of the coating systems:

Clearcoat Basecoat Primer Electrocoat Phosphate METAL Phosphate Electrocoat

Clearcoat Basecoat Primer PLASTIC

It is easy to predict that the clearcoat must be designed to withstand environmental impact (effect of parameters of performance). For this reason, the emphasis is given to the clearcoat in this report. The general literature l lists currently used clearcoat systems, which include: one-component acrylic-melamine, one-component polyurethane, two-component polyurethane, one-component waterborne, one-component powder. Powder coatings are still on the stage of development and extensive testing thus some data should be updated in future.

Automotive Clearcoats

297

The weathering performance (durability) depends on the chemical composition which must include all components of the mixture since each component, even used in very small quantity, may contribute to photochen1ical degradation. In order to describe composition, recent patents 2- 14 obtained by the major manufacturers of these materials were analyzed to construct a list of individual cOlnponents given below. Components of automotive clearcoats: Polymeric materials:

Powder coatings: 47,48,55,58 • copolymer of methacrylate, ll1ethyl & butyl methacrylate, and styrene with epoxy functionality cured with diacid or uretdione (HDI, IPDI) • polyacrylate polyol (MMA, esters of acrylic & methacrylic acid, styrene) OH group functionality polyester polyol (dialcohol + diacid) cured with aliphatic or cycloaliphatic ketone (ketoxill1e) polyisocyanate or isocyanurate • polyester (hydroxymethacrylate, n-butylacrylate, MMA, neopentyl glycol, and dicarboxylic acid) with OH functionality, polyacrylate containing hydroxyl group cured with HMDI blocked with 1,2,4-triazole • acrylic copolymer (styrene, methacrylic acid, butyl & methyl methacrylates cured with crosslinker of carboxylic groups (epoxides or oxazolines) Solve 11 t-conta i11 i11g: 47-54,56,57,59,60 • acrylic resin OH terminated alkoxysilyl group-containing copolymer (urethane or siloxane bonding) cured by reaction of hydroxyl group from acrylic resin with alkoxysilyl • acrylic resin with OH functional groups cured with aminoplast (condensate of formaldehyde and urea, thiourea or melamine) Resimene 755 from Monsanto or Cymel 1130 (methylate melan1ine-foffi1aldehyde cond.) • acrylic polymer with OH groups microgel based on acrylic cured with aminoplast or polyisocyanate (2-colnponent system) • organosilane polymer (styrene, methacryloxy propyltrimethoxy silane, and trin1ethylcyclohexyl n1ethacrylate) acrylic polyol (styrene, alkyl methacrylate, hydroxy alkyl acrylate) - macrogel urea by reaction of Resimene 755 from Monsanto or Cyn1el 1130 (methylate melamine-foffi1aldehyde cond.) • polyol (caprolactone copolymerized with 1,4-cyclohexanedimethanol) star polymer (ehyleneglycol dimethacrylate, methyle, benzyl and 2-hydroxyethyl methacrylates) cured with isocyanurate or aminoplast (Cymel 1133) • acrylic polymer (styrene, alkyl methacrylate, hydroxy alkyl methacrylate) with OH functionality polyol (caprolactone copolymerized with 1,4-cyclohexanedimethanol) cured with isocyanate (triphenyhnethane triisocyanate or trimer of hexamethylene diisocyanate

298

Weathering of Plastics

• acrylic polymer (styrene, MMA, alkyl methacrylate, alkyl acrylate) crosslinking acrylic (the same but containing glycidyl) • acrylic resin aminoplast (Cymel 1130) • acrylic polymer (hydroxypropyl acrylate, styrene, butyl acrylate, butyl methacrylate, acrylic acid) cured with aminoplast (CymeI 1130) In SUlTIlnary, the following polymeric materials will be analyzed in the section discussing chemical mechanisms of degradation: • acrylic polymers and copolymers • polyurethanes • aminoplasts The importance of this analysis is to include typical chemical groups in order to predict potential products of degradation.

Solvents • • • • • • • • • • • • •

xylene Solvesso 100 n1ethanol butanol, iso-butanol mineral spirits heptane butyl acetate ethyl acetate methyl ethyl ketone acetone dipropyleneglycol monomethylether methyl amyl ketone hexyl acetate

Initiators various initiators used in polymerization of acrylic resins

UV stabilizers • HALS (Tinuvin 144,292) • UV absorbers (Tinuvin 400, 900, 1130)

Catalysts • tin (most frequently DBTL) • amIne

299

Automotive Clearcoats

Flow/rheology modifiers • Perenol F30 and F45 - polyacrylates • Modaflow PIlI - polybutyl acrylate • polydimethyl siloxane oil • Byk 361, 323 & 325 - polyacrylates • BYK 306 - polyether modified dimethyl polysiloxane Other components • fume silica • phosphites

EFFECTS OF PROCESSING Processing effects are given in Table 4.

Table 4. Process parameters, their potential effects, and induced modes of failure. Process parameter Altered composition Production in spring and sumnler

Potential effect Induced mode of failure durability, Quality of finish all modes of failure possible increased acid etch 25 which can be compensated by cracking exposure to UV under dry, cool conditions delamination mar

Reduced rotation speed of spraying random orientation of metal flakes, orange pee1 28 ,36 bells popping, fuzziness, wrinkling, poor gloss28

Residual moisture in the basecoat Dust in plant3o ,43

cracking

craters; cars need to be repainted with different paint potential corrosion (more initiator) faster degradation

Higher temperature of baking 30 Lower film thickness

lower durability cracking

degradation products

chromophores

in solvent-base paints shorter life, in powder paints corrosion uneven finish (particle size too close to filnl cracking thickness)

Particle size 35

surface defects

cracking

HUlnidity43

gloss (lower durability)

delanlination

mar Spray gun orientation

Paint volume output vs. line speed

thickness uniformity

cracking

thickness

corrosion cracking

300

Weathering of Plastics

NUlnerous effects can be induced by the method of processing and precision of equipment operation. At the same time, it should be considered that probabilities of these inconsistencies in production are very low because automotive companies have invested in very sophisticated equiplnent which prevents such artifacts. It is very essential to note that many of these failures are related to film uniformity and that film uniformity can completely change coating perforn1ance in relationship to its durability.61 These effects are discussed further in the next section. Similarly, errors in composition may seem very remote since paint manufacturers are very experienced. At the same time, present coatings (clear coat/base coat) are very risky in automotive applications because of their weathering properties. Previously used coatings deteriorated in a gradual process initiated by a loss of gloss. It was therefore possible to obtain early waluing that particular paint (batch) does not work. In the case of lnodern paints, this warning does not exist, only catastrophic failure (cracking, peeling) suddenly occurs without much detectable difference in perfonnance prior to the failure. Under these circ*mstances, precise control of coatings prior to their application makes good business practice, considering that in-field failure is very expensive.

MECHANISMS OF FAILURE Many aspects of degradation must be analyzed to reach expected understanding which allows one to pinpoint chemical changes contributing to the modes of failure included in the Table 3 and to find candidate n1ethods which can predict failure. Some of these data can be found in the existing literature l ,17-26,46,61-76 and some mechanisms are still not fully understood. First, we need to analyze the mechanisms of degradation of individual polymers which are used for the production of clear coats as listed above. These polymers include: acrylic polymers and copolymers, polyurethanes, and aminoplasts. The analysis is performed to select the most important reactions which determine durability of automotive coatings. Figure 1 shows typical reactions of acrylic resins. These reactions are ituportant for all three types of resins used in automotive clear coats because they all contain acrylic backbone but differ in the method of chain extension (cure). Acrylic resins are UV stable. They are only degraded because of presence of photoinitiators fron1 impurities. The initial step of photochemical degradation consists of macroradical formation. This first step opens numerous possibilities such as chain scission, crosslinking, formation of hydroperoxides, and formation of carbonyl groups. It is impoliant to mention that there is a general agreement that these changes take place but the kinetics of these changes varies. For example, one research group presents data indicating decrease in carbonyl group formation. 73 In other paper,77 there is an experimental evidence of a reverse trend. This information is very essential to follow degradation because

Automotive Clearcoats

301

Formation of radicals CH3

CH3

CH,- C - CH, -

COOH

This reaction may lead to formation of hydroperoxides \vhich may

COOH

either decompose producing carbonyl groups or becollle precursors of further chain scission. The examples below show two reactions CH3

CH3

-CH1-~-CHz-fCOOH

that affect molecular weight:

CH3

I I -CH2-~-CH2-1-

+ CH3

+ t:OOH

COOH

chain scission (molecular weight reduction): -CH.,-~H-CH.,tO~H

-----+

-C H2 - CH =C H2 +

-r.

COOH

crosslinking (molecular weight increase)

Formation ofmacroradicals and subsequent reactions affecting molecular weight occur also (in similar sequence of reactions) in esters: -CH.,-CH-

COOR

-eo I

-CH2-CH-

The above reactions give exarnples offomlation ofmacroradicals which only occur due to abstraction ofa side group with a help ofphotoinitiators rather than by a direct action ofUV itself(bonds involved are UV stable to sunlight radiation). These reactions also show that carbonyl groups are lost in the process of photolytic degradation, although they can be also formed from decomposition of hydroperoxide as represented by the following reaction:

-CH.,-CH-

OOH

Figure 1. Typical reactions of acrylic resins.

302

Weathering of Plastics

this is one of the basic measurements. Since there is an agreement among most research groups that correlation (between, for example, natural and laboratory exposures) requires verification ofmechanism, it is difficult to reconcile this statement with the fact that such different estimations of fundamental product of degradation exist. More comments on this subject are included below. Hydroperoxide concentration depends on two competing reactions: oxidation of macroradicals and decomposition of hydroperoxide by heat or UV. Here is one important parameter of weathering - temperature - which plays an essential role in the studies of these materials. Depending on temperature, reaction may take different course. Figure 2 shows another potential anomaly in the course of degradation which occurs when wrong wavelength of light is used in the studies. During such reaction main chain scission occurs which never happens in the outdoor environment. CH3 I

CH3 I

-CH2-C-CH2-CI

0= COCH3

-hv -.

COOCH3

Figure 2. Reaction not typical of outdoor exposures of acrylic resins.

In polyurethane clear coats, urethane linkages are formed. Figure 3 shows two potential reactions which may take place at urethane linkage. Both reactions have low probability which is most likely the most important reason for which urethane coatings are used more frequently than aminoplasts. Especially, in regard to acid etching, polyurethanes are superior to other coatings which have either ether or ester linkages. 26 Cleavage ofC-N bond: -O-C-NHII

---+

o-cII a

Hydrolysis:

Figure 3. Typical reactions of urethane bonding. 62

The mechanism of acid etching of melamine cured systems is given in Figure 4. Presence of water and acid causes hydrolysis of ether linkage which changes molecular weight and thus physico-mechanical properties of coating. 26

Automotive Clearcoats

303

HO OH

Figure 4. The mechanism of acid etching. 26

These changes prompted some research groups20,23,25 to conduct extended studies especially in connection with field observations that cars produced during fall or winter have more durable paint than those produced in spring and summer. Figure 5 explains perceived mechanism. If car is painted in winter, the coating cures at dry, cool conditions which ultimately leads to the last compound to the left in the 2 nd row. These changes do not cause a change in molecular weight ofpolymer forming coat. If the hydroperoxide (compound at the right ofthe 2 nd row) is decomposed by UV or heat then changes eventually lead to hydrolysis which weakens coating (last formula at the left of the 3rd row). Similar coating protection can be achieved by controlled exposure of coating to UV. The proposed mechanism helps to understand some problems with melamine coatings. In addition, it indicates importance of other parameters of weathering such as temperature, humidity, and acid rain. In summary, one may observe that some progress was n1ade in qualitative understanding ofchemistry ofautoillotive degradation. At the same time, there is still deficiency in quantitative data - necessary to control mechanisms during an experiment (outdoor, laboratory, or correlation of both).

INTERRELATIONS BETWEEN THE PERFORMANCE CO-NDITIONS, THE MODES OF FAILURE, AND THE CHEMICAL MECHANISMS OF DEGRADATION Table 5 lists these interrelationships for the modes offailure from Table 3, typical parameters of performance from Table 2, and information included in the literature on the mechanisms of chemical degradation in relationship to failure modes.

Weathering of Plastics

304

/ N

N.J-. N

\ I,.. II N~N~N

)

\ N

N .... H

N~N

)

/ N

~ ·,H N

\A)l

,,0·

0) \

! this branch applies to \vinter production

this branch applies to summer production Figure 5. Photooxidation mechanisms ofmelamine. 25

hv or heat

Automotive Clearcoats

305

Table 5. Mode of failure versus parameters involved and chemistry of changes. Mode of failure

Parameters involved

Chemical changes involved

Gloss

UV radiation

loss of amide and urethane (PU), loss of ether and triazine not well resolved (n1elamine),23 carbonyl decrease/ 3 carbonyl increase and crosslink scission correlates with hydroperoxide concentration,77

irradiance level

increased irradiance does not always accelerates degradation

temperature

increase in carbonyl & decrease in triazine on temp. increase by lO°C 23

hun1idity

melamine photooxidation rate increase with humidity increasing l8

sulphuric acid shrinkage

loss of anlide (PU), loss triazine (n1elanline)23 loss of volatiles 23

Yellowing

UV radiation temperature

no specific data

Adhesion loss

UV radiation telTIperature moisture pH

oxidation of lower layer (basecoat),18

Cracking

UV radiation temperature

Mar

generally related to photooxidation but no data and correlation with gloss decrease 24 no specific data no specific data see UV radiation

UV radiation

affects crosslinking loss (no specific data)

precipitation (pH)

accumulation of dust particles helps to retain moisture and acid 26 car washing resistance correlates with Taber test 26 and stress-strain19 concentration of water in film depends on hydrophobicity of film 26

acid rain (pH) particle embedding lTIoisture UV radiation

Surface erosion

no specific data no specific data accelerated by cOITlbined action of UV and pH (decreasing) 17

moisture pH

abrasion H 20 absorbed temperature Water spots

no specific data; affected by weathering equipment (QUV different than W_O_M)74

increases water penneability; coating may have higher temp. than T pH affects surface etching rate,17,26 several acids in composition17 no specific data no specific data effect confirnled, 17 no specific data

UV radiation

effect confirmed, 17 no specific data

oxygen

no specific data increases with pH decreasing, 17,26 new paint mostly affected, 1 no specific data

dissolved acids pancreatine

Oil staining

used oil

staining due to carbonaceous materials,38 no specific data

Substrate corrosion

deicing salt

n10st severe cases are due to the loss of environn1ental protection due to the damage of coating: mechanical or photochemical 46

saltwater particles

306

Weathering of Plastics

There are numerous publications available which deal with the subject (36 publications references during 2 two years) and extensive infonnation available on qualitative reasons for automotive coating degradation. The quantitative data are still very scarce. From the above list, it is easy to note that only a few chemical changes can be selected as a base for quantitative measuren1ent of the degradation rate (based on existing data). Gloss change is the most investigated mode offailure and perhaps there is a possibility to select methods of chetnical analysis which may correlate with gloss. At the same time, experts 18 ,24 in the field clearly indicate that gloss loss is not the major problem of clear coat/base coat systems. Moreover, it is indicated24 that good gloss retention cannot preclude catastrophic failure of coating which occurs by peeling and cracking. Frequently, these last two failures are described as "unpredictable". This seems to signalize the nature of the problem of the lack of correlation which is discussed in more detail below. Sitnilar systems are used for coating plastic parts of an automobile. They also include clear coat/base coat system. Several current publications deal with this subject. 42 ,65,67,68 Two directions are taken into consideration: development of directly paintable and adherable polyolefin compounds and preparation of TPO for painting. If the first direction prevails in future (more novel solution) then weathering aspects will be described by similar relationships. If the second n1ethod prevails then preparation method of a surface must be included. These methods include: chemical oxidation, corona discharge, flame treatment, plasma treatment, UV/benzophenone surface degradation, and adhesion promoters. Except for the last method, all methods used affect photodegradation since all these methods induce potential defects which may initiate further degradation which must be accounted for. Finally, the above discussion included only chemistry of degradation. At the same time, it is well known that the structure offilm (unevenness, defects, orange peel, problems offlow, problems with sintering ofpowder coatings) has essential bearing on its degradation. There is no data to report on this matter and thus there is a clear need for extensive research in this area, considering that initial defects in the film surface can alone ruin chances to obtain correlation in experimental work.

SPECIMEN TESTING Some existing standards define testing method which should be used for the evaluation ofexposed specimens. These methods include: • description of changes to appearance7, 11,13 • comparison with origina1 7 • testing according to material specification5 • color change 3,4,13 • glossll,l3

Automotive Clearcoats

307

• physical properties 11 • mechanical properties] 1,13 The above methods are important for the final product evaluation but do not have any predictive value which can be used in the design of weathering experitnent which may help to establish correlation. There are several analytic methods which are used to follow a degradation rate: • FTIR to determine carbonyl, melamine crosslink density, and amide II in PU 77 ,78 • photoinitiation rate (PIR) based on ESR measurement 79 • hydroperoxide tiration80 • surface composition by XPS 39 • orange peel by image analysis 37 All the above methods are suitable and they can eventually contribute to the selection of laboratory weathering conditions. At the same time, the methods have some ilnportant deficiencies. ESR measurement is time-consuming and expensive. FTIR and XPS methods are affected by the surface contamination of a specitnen which is especially important in outdoor exposures. Carbonyl determination does not allow to distinguish between carbonyl loss, due to degradation of carboxyl and ester groups, and carbonyl gain due to hydroperoxide decomposition. Hydroperoxide titration gives reliable data but there are always two competing processes during degradation: hydroperoxide formation and hydroperoxide decomposition. It is therefore difficult to determine extent of photochemical reaction. Fronl the above, a clear need for a further search of chemical analytic method is needed. In addition to the study of selective chemical change there is a need to assess distribution of changes. The so-called "catastrophic (without wanling) failure" clearly indicates that a part of a mechanism of cracking must be related to the changes in crystalline structure which Inakes material increasingly non-uniform to cause sudden (unexpected) crack. Several other opportunities still exist for monitoring the degradation. One is described in the separate chapter of this book. Clearcoats retain their properties due to a high addition of UV stabilizers. Therefore the method of monitoring the concentration of active stabilizers is another useful approach. Recent developments in image analysis allow for simultaneous monitoring of gloss, color change, and surface changes such as formation of haze, microcracking, delamination, etc. This methods tested for sealants applications 81 have proven to be very efficient in durability assessment. If there is a choice between direct determination of defects and indirect (such as chemical analysis) the direct method should always be selected since it provides information on changes directly responsible for perceived failure. The chemical analysis is still very useful because it allows to confirm mechanisms ofchange - useful in remediation of the problem.

308

Weathering of Plastics

EXPECTED LIFETIME For the lack of cOlTelation with studies conducted in laboratory the only requirement used for OEM coatings is that of 5 years Florida exposure without failure. All other methods are still auxiliary techniques used more to accumulate data and experience than as a screening procedure. There is a clear need to develop an expected lifetime in Xenon-arc Weather-a-Meter and EMMAQUA, even ifbased on energy assumption as a starting point. If such standard is not clearly stated (and results not compared with) false expectations regarding laboratory exposures will always exist. Bauer21 recently suggested a new approach to the prediction of durability of a painted car and in view of these considerations such standard is essential.

NATURAL EXPOSURE The precise guidelines can be found in SAE standard. 14 Coating systems are exposed in the panel exposure racks and black boxes. Alten1ative method of outdoor exposure includes the use of solar Fresnel reflector apparatus. 15 Environmental data include: total solar radiation, total UV, optionally selected wavelength radiation, and time of wetness. In Fresnel reflector exposures, it is necessary to determine radiant exposure, elapsed exposure time, black panel temperature, and spray cycle.

LABORATORY EXPOSURE The summary of standardized laboratory methods of exposure is given in Table 1. It can be additionally mentioned that there is an interest in extending laboratory methods to include the effect of acid rain on weathered coatings. Interesting modification of SAE J1960 is re23 ported. Panels were removed for 1h three times a week and sprayed with solution (pH=3.2) of mixture of sulphuric, nitric, and hydrochloric acids in proportions 1:0.3 :0.17.

CORRELATIONS The situation is well characterized by two statements included in Bauer's paper: 18 "Given the complex photodegradation chemistries that occur in these coating systems, a lack of correlation between outdoor exposures and conventional accelerated tests, which employ harsh exposure conditions, is not surprising."

Automotive Clearcoats

309

"It is clear that the predicting free radical photooxidation requires measurement ofboth K (reaction constant) and hydroperoxide concentration." These two statements include several important messages: • harsh exposure conditions • complex photodegradation chemistry • n1easurement Further discussion concentrates on these subjects. It is absolutely certain that the industry needs to accelerate testing. Otherwise, product improvement will be slow. There are two options which can be exploited to achieve this goal: • increase values of quantities involved in photodegradation • find "magnifying glass" The first option was tried for many years. Various equipments and sets of parameters were tested and, since correlation is still not available, failed. Most researchers in the field of durability of materials agree today that acceleration of testing cannot be done by modifying test environment. Also, the reason is clear - complex photodegradation chemistry does not allow to predict what such changes in parameters will affect. It is thus clear that one has to simulate in laboratory conditions typical of natural environment. There is no particular barrier in equipment which would not allow to achieve consistent control of • radiation wavelength • radiation intensity • tenlperature (composite of air temperature, infrared, and specimen color) • humidity The above are the main parameters controlling photodegradation and they can be controlled with precision (see two other chapter in this book on application of different equipment to studies of automotive coatings). The current developments in weathering devices allow one to run any complex program, such as for example, close simulation of seasonal effects. There are two environmental parameters which are not currently simulated in weathering devices. These are stress and pollutants but at the same time there are many methods to include them using the existing equipment (one example was discussed in the previous section). The key to the further development is to find nlethods which allow to verify if the chosen program of weathering conditions allows to follow degradation in the outdoor environnlent. In order to achieve this, the future work should concentrate on the understanding of degradation mechanisms rather than looking for universal new machines for testing as discussed in one recent publication. 24

310

Weathering of Plastics

In order to use the second method (nicknamed "magnifying glass") designed to shorten testing time (or tinle of decision point), two directions of studies are needed: • understanding a chemical mechanism of degradation • establishing a consistent indicator of degradation useful in measuring kinetics of degradation. There were made comments on this subject in the recent paper: 24 "There is little reason to suspect that comparable composition changes should have comparable physical repercussions in coatings from different chemical faluilies." "Weathering tests based on chemical composition change rates provide no information about the physical repercussions of the chemical changes. Therefore, these tests can make no comment on the physical tolerance of clearcoats to the chemical composition changes they undergo, leading to possible erroneous conclusions regarding their durability." The above comments suggest that the fact of detecting a certain concentration of, for example, carbonyl groups does not nlean that a coating, regardless of its formulation, is bound to fail. At the same time, it is possible to observe that the particular coating fails when it attains a particular composition of carbonyl groups, providing that the conditions of degradation (determining the mechanisms of photochemical changes) where the same. This sets the goals for experiment design which may offer correlation: • prior to the experiment the chemical mechanisms of degradation were sufficiently understood to select a measurable quantity which allows to check that mechanisms of degradation are the same in two correlated environments • physical parameters are chosen to have close proximity of exposure conditions • a measurable quantity allows to detect early changes which have been found to signalize particular failure. In automotive coatings, this stage was not reached yet in spite of extensive effort. One reason is that, in most studies, goals too difficult to achieve were set. In the most extensive studies, attempts were made to find universal method, whereas there are no universal changes for, say, polyurethanes and melamine cured acrylics. On a surface, they produce the same hydroperoxides, carbonyls but these concentrations "can make no comment on physical tolerance of clearcoats". There are many examples in automotive coatings which show that focusing on a particular problem helps to solve it. When difference between summer and winter products was observed, the problem was solved as described in Figure 5. When the hydrolysis of aminoplasts was discovered as described by Figure 4, polyurethane coatings gained markets. Many years ago, coatings were degrading because some undesirable solvents

Automotive Clearcoats

311

were used which where then eliminated. UV stabilizers were not giving protection and this problem was eliminated because new stabilizers were introduced to assure their lower volatility during coating baking. The first powder coating was developed long time ag0 55 and it did not perform because today's rheological additives and UV stabilizers were not available to support idea. Present powder coatings are close to the required perfonnance. All these examples show that well focused effort can produce results. The second reason can be related to very rapid changes in automotive coatings which did not allow to stabilize situation. Before any required mechanisms were found, a new range of products was introduced and work had to be repeated. The third reason is related to the fact that too many exploratory research works were needed to scrutinize the test n1ethods which can be useful. It seems that it is a matter of time when proper correlations will be developed. In order for this to happen, the approach must include fundamental analysis of the problem which can be narrowed down to • exposure should simulate conditions found in environment of material performance • verification of these conditions should be established by the use analytical factor which confirms that chemical changes are the same in compared studies • the modes of failures of interest should be related to the chemical changes which can be easily measured.

REFERENCES 1.

2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16.

A. Jurgetz, ~o/fetal Finishing, 93, 10, 53-5 (1995). Plastic tnaterials and coatings for use in or optical parts such as lenses and reflectors of high-intensity discharge forward lighting devices used in motor vehicles, SAE J1647 MAR95. Accelerated exposure of automotive exterior materials using a controlled irradiance water-cooled xenon arc apparatus, SAE J1960 JUN89. Accelerated exposure ofautonlotive exterior nlaterials using a controlled irradiance air-cooled xenon arc apparatus, SAE J2019 JAN94. Accelerated exposure of autonlotive exterior materials using a fluorescent UV and condensation apparatus, SAE J2020 MAY95. Testing the crack resistance of clear coats in two-coat nletallic finishes, VOA 621-4, German Motor Manufacturer's Association. Weather-Ometer test, LP-463 PB-16-0 1, Chrysler Corporation. Condensing humidity resistance test, LP-463PB-9-0 1, Chrysler Corporation. Resistance to artificial weathering, BO 101-1 , Ford. Procedures for laboratory accelerated exposure of automotive tnaterials, GM9125P, General Motors. Weatherability and light resistance test methods for synthetic resin parts, MO 135: 1990, Nissan. Methods of testing for motor vehicle paints. Part 12: Resistance to accelerated weathering. BS AU 148: 1969. Test nlethod of weatherability for automotive parts, JIS 00205. Outdoor weathering of exterior tnaterials, SAE J1976 FEB94. Accelerated exposure of automotive exterior materials using a solar Fresnel Reflector apparatus, SAE J 1961 JUN94. W.J. Putman, M. McGreer, and O. Pekara, STP 1294, ASTM, 1996.

312

17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69.

Weathering of Plastics

U. Schulz and and P. Trubiroha, STP 1294, ASTM, 1996. D. Bauer, J. Coat. Techno!., 66,835,57 (1994). 1.L. Courter, J Coat. Techno!., 69,866,57 (1997). K.M. Wemstahl and B. Carlsson, J Coat. Technol., 69, 865, 69 (1997). D. Bauer, J Coat. Tec/mol., 69,864,85 (1997). 1.H. Braun and D.P. Cobranchi, J. Coat. Technol., 67, 851, 55 (1995). K.M. Wernstahl, Po(vm. Deg. Stab., 54, 57 (1996). M.E. Nichols, 1.L. Gerlock, and C.A. Slnith, PO~l'm. Deg. Stab., 56, 81 (1997). P.H. Lamers, B.K. Johnston, and W.H. Tyger, Polym. Deg. Stab., 55, 309 (1997). B.V. Gregorovich and 1. Hazan, Prog. Org. Coat., 24, 131 (1994). U. Pitture Vernici, 71, 20, 16 (1995). T. Triplett, Ind. Paint POlvdel~ 71, 7, 20 (1995). E.A. Praschan, ASTM Standardization News, 10,40, 1995. J.M. Bailey, Ind. Paint Powdel~ 71, 7, 14 (1995). 1. Schrantz, Ind. Finishing, 67, 10, 20 (1991). J.M. Bailey, Ind. Finishing, 67, 4, 30 (1991). R. Jaeger and S. Kernaghan, Obeljlaeche/JOT, 35, 9, 42 (1995). Anon., Obeljlaeche/JOT, 36,9, 16 (1996). J.M. Bailey, Ind. Paint POlvder, 70, 12, 10 (1994). J.T. Guthrie and A.P. Weakley, Sillf Coat., Int., 79, 2, 58 (1996). R.T. Quazi, S.N. Bhattacharya, E. Kosior, and R.A. Shanks, Sillf Coat. Int., 79, 2, 63 (1996). H. Schmidt and D. Fink, Sillf Coat. Int., 79, 2, 66 (1996). T.E Barton, D.C.W. Siew, and S.E. Werner, Slllf Coat. Australia, 33, 4,18 (1996). H. Schmidt, Paint Ink Int., 9, 3, 1994. S.L. Kiefer, Paint Ink Int., 8, 12 (1995). E. Lau and D. Edge, ANTEC'93, 2487. B.A. Graves, Products Finishing, 59, 10,48 (1995). B.A. Graves, Products Finishing, 55, 10, 42 (1991). U. Biskup, PUture Vernici, 72, 1, 13 (1996). A. Amirundin and D. Thierry, Prog. Org. Coat., 28, 59 (1996). US Pat. 5,508,337, Bayer Aktiengesellschaft, Gennany, 1996 US Pat. 5,492,955, Bayer Aktiengesellschaft, Gennany, 1996. US Pat. 5,283,084, BASF Corp., USA, 1994. US Pat. 5,354,797, E.!. Du Pont de Nemours and Company, USA, 1994. US Pat. 5,244,696, E.!. Du Pont de Nemours and Company, USA, 1993. US Pat. 5,159,047, E.!. Du Pont de Nen10urs and Company, USA, 1992. US Pat. 4,937,281, E.!. Du Pont de Nemours and Company, USA, 1990. US Pat. 4,402,983, E.r. Du Pont de Nemours and Company, USA, 1983. US Pat. 4,808,656, PPG Industries, Inc., USA, 1989. US Pat. 4,680,204, PPG Industries, Inc., USA, 1987. US Pat. 5,580,660, DSM N.V., Netherlands, 1996. US Pat. 4,728,543, Nippon Paint, Co. Ltd., Japan, 1988. US Pat. 5,063,114, Kanegafuchi Kagaku Kogyo Kabushiki Kaisha, Japan, 1991. M.A. Grolitzer and D.E. Erickson, Waterborne, Higher-solids, and Powder Coatings Symposium, 1994. G. Wypych, Handbook of l\laterial Weathering, 2nd edition, Chem Tec Publishing, Toronto, 1995. U. Schultz and R.-D. Schulze, Oberjlaeche/JOT, 35, 9, 62 (1995). K. Gaszner, M. Heinrich, and T. Schuler, Aluminum, 71, 6, 751 (1995). M. Hartung, H. Hintze-Bruening, and H.-J. Oslowski, Metallobeljlaeche, 50, 6,494 (1996). T. Suzuki, T. Tsujita, and S. Okamoto, Eur. Coat. J, 3, 118 (1996). C. Daniels, Products Finishing, 56, 11,64 (1992). A.S. Wimolkiatisak, A.S. Scheibelhoffer, D. Chundury, and P.M. Mokay, ANTEC'92, 296. L.W. Hill, H.M. Korzeniowski, M. Ojunga-Andrew, and R.C. Wilson, Prog. Org. Coat., 24, 147 (1994).

Automotive Clearcoats 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81.

313

G. Linger and E. Hess, S10! Coat. Int., 79, 2, 66 (1996). lL. Gerlock, W. Tang, t\-1.A. Dearth, and TJ. Komiski, Polym. Deg. Stab., 48,121 (1995). J.L. Gerlock, T.J. Prater, S.L. Kaberline, and J.E. deVries, Polym. Deg. Stab., 47,405 (1995). N.S. Allen, MJ. Parker, CJ. Regan, R.B. McInture, and W.A.E. Dunk, Polym. Deg. Stab., 47, 117 (1995). C. Gopsill and P.W. Griggs, Sla! Coat. Int., 76, 6,247 (1993). D.S. Allan, N.L. Maecker, D.B. Priddy, and NJ. Schrock, Macromolecules, 27, 7621 (1994). B.L. Rytov, V.B. Ivanov, V.V. Ivanov, and V.M. Anisimov, Polymel~ 37, 25, 5695 (1996). D.R. Bauer, D.F. Mielewski, and lL. Gerlock, Polym. Deg. Stab., 38,57 (1992). D.R. Bauer, lL. Gerlock, and D.F. Mielewski, Po~vnl. Deg. Stab., 27, 271 (1990). lL. Gerlock, D.F. Mielewski, and D.R. Bauer, Polym. Deg. Stab., 20,123 (1988). D.F. Mielewski, D.R. Bauer, and lL. Gerlock, Polym. Deg. Stab., 33, 93 (1991). G. Wypych, F. Lee, B. Pourdeyhimi, Comparative study of sealants durability. Surface changes, RILEM Symposium 2000.

Index

A ABS 61 accelerated electrons 178 accelerated tests 10 acceleration factor 12 acid etching 303 acid rain 162 acids 218 acrylic-melamine 296 acrylics 297 activation energy 173 additives 253 agriculture 218, 225 alkalinity 217 alkyd-lnelalnine 293 alninoplasts 298 alnorphous 77, 183, 199 antagonistic interaction 225 antioxidants 169, 179, 225, 234 antistatics 5 appliances 99 Arizona 5, 19,72,77,93 Arrhenius activation energy 169 equation 171 plot 180 ASA61 AIR 78,186 autoillotive 29, 43, 72, 161, 185, 241 autolnotive coatings 293

B bags 211 basecoats 293

benchtop instrulnents 10 bioburden 178 biocides 5 blends 212 blistering 151 bond breaking energy 62 cleavage 70 bottles 211 Brabender 212 branching 81 brittle layer 178 brittleness 63 buildings 15, 70, 133

C calciuln stearate 84 carbon black 127 carbon fiber 99 carbon-arc 7, 16 carbonyl groups 78, 228, 261, 301 catastrophic failure 307 chain cleavage 78, 215 scission 141, 170, 301 chenlical resistance 237 chelni-crystallization 141, 149 chemiluminescence 170 chromophores 97, 228 ClRA8 clearcoat 185, 296 clilnates 17, 261 coatings 186 coextrusion 93, 271 color 61, 96, 134

Index

316 compatibilization 212 cOlnpressive stress 155 COlnpton secondary electrons 178 condensation 5 conductive cooling 137 construction 161 Inaterials 1 consumer goods 161 containers 211 corrosion 151 CPE 61 CPVC 61 crack depth 182 cracking 151 cracks 102 crosslinking 301 crosslinks 267 crystalline regions 199 crystallinity 141, 197, 214 change 142 crystallite size 197 crystallites 170 crystallization temperature 213

D dark colors 138 daylight irradiance 4 debonding 102 degradation mechanism 10 - 11 rate 1 telnperature 214 degrading parameters 11 delalnination 151, 307 design life 2 diffraction 197 diffusion 229, 253 DIN 113 discoloration 162 disposal 2 DSC 169, 177

durability 169 durability testing 2

E early degradation 218 elongation 70, 230, 233 elnbrittlelnent 173, 233 end-groups 77 EPDM 170 epoxy resin 99 EPR 170,178 equatorial tracking 24 equipment 7, 50, 105 ESR 186,307 ethylene-propylene CopolYlner 262

F factors of aging 69 failure 1, 2, 178, 188, 295 criteria 137 fiber 233 fiber-reinforced plastics 99 fibrils 173, 182 films 180, 218, 225 filter system 34 flaking 151 flame retardants 71, 238 Florida 5, 18 - 19, 72, 93, 236, 273, 308 fluorescent lamps 7, 16 Fresnel-reflecting mirrors 16 FTIR83, 186,212,229,286 fuel combustion products 162

G gas fading 162, 248 Gaussian distribution 199 GC 282 glass fiber 103, 195 glass transition 183 gloss 306 GPe 78,94,212,263

317

Index

greenhouses 225, 271

H HALS 161, 164, 185,225,233 deactivation 222 haze 96,307 heat buildup 133 deflection telnperature 136 HIPS 61 HMDI297 hot water 99 HPLC 177 hUlnidity 309 hydrolysis 5, 61 hydroperoxide titration 186 hydroperoxides 170,227,268, 300

I imnlersion 155 tests 99 il11pact strength 61, 233 inductive coupled plasllla 84 infrared energy 4 heating 121 initiation 227 injection molding 141 installation 2 interface 103, 170 international organizations 15 IPDI297 IR 107 reflective piglnents 134 irradiance 3 - 4, 47, 95 irradiation 105 ISO 113 isocyanurate 297

L latitude 22

layer rellloval procedure 141 lead stabilizer 277 life prediction 15 light penetration 203 lightfastness 29 long-tenn data 10 Lorentz 199 low-volatility 272

M macroradicals 301 maintenance 2 marine coatings 151 Inaterial degradation paralneters 2 mathematical weighing process 135 Inatrix 102 Ineasureillent geometry 135 Inechanical stress 83 melaluine 304 Inelting telnperature 213 nletallocene 69 methane cOlnbustion 162 luethods of measureillent 12 microcracking 307 Inicrocracks 233, 274 migration 72 milling 262 111iniblinds 277 M n 79 l110de of failure 306 lTIodulus 141 lTIoisture 5 monoclinic fonl1 207 Mz/M n 81

N nanoconlposites 196 NMR 186,212 Norrish 206 NO x 162

318

o OIT 69, 169, 177 optical properties 134 outdoor exposure 15 oxidation 78 oxygen 261 diffusion 182

p paint 185 parafocus 199 PAS 186 performance criteria 10 - 11 pesticides 225 pH 177 phenolic antioxidants 161 photohydrolysis 97 photoinitiators 300 photons 178 photo-oxidation 141 pigtnents 127, 134,237, 265 pKa-value 217 plasticizers 5, 286 plastics 1 PMMA61 polarization 199 polishing 200 pollutants 5 polyacrylate 297 polyalnide 5, 61 polybutyleneterephthalate 172, 195 polycarbonate 3, 5, 271 polyester 5, 170, 297 polyethers 61 polyethylene 61, 170, 178, 211, 218, 225, 261, 286 polyethyleneterephthalate 77, 93 polyol297 polyolefins 211, 21 7 polyoxymethylene 99, 129 polyphenylene sulfide 282

Index

polyphenyleneether 99 polyphenylenesulfide 99 polypropylene 61, 71, 144, 170, 177, 211, 233, 261 polystyrene 56, 61 polyurethane 39, 296 polyvinylchloride 61, 83, 107, 277, 286 post application 2 post-irradiation 169 failure 184 powder coating 296 prediction tnethods 174 preliminary studies 10, 13 product temperature 4 propagation 182 pUlnps 99

Q quenchers 70 QUV 84, 262, 277

R radiant energy 133 radiation 7, 177 global 107 simulation 46 radical trapping 71 radicals 77, 170, 181, 231, 268 Ralnan 186 reaction kinetics 108 reactive chetnicals 218 recycling 211 reference Inaterial 35 reflectance 135 relaxation 141 relief melnbrane 283 renewable resources 2 replacen1ent 2 residual stress 141,147 - 148, 151 rigid vinyl 136 routine testing 11

319

Index

S salt formation 218 scattering 199 scission 263 secondary crystallization 141 SEM 99, 173, 182,284 sensors 99 shelf-life 169 shrinkage 151 siding 83 skylights 271 slnog 162 solar absorptance 136 solar cut-on 3 - 4 solar spectfUln 133 - 134 solidification 141 solubility 253 spectral irradiance 106 spectfUln 8 spinning stability 221 stabilization 70 stabilizers 233, 267 package 225 standard reference Inaterial43 standards 49 sterilization 177 strain 143 stress 6, 143 build up 154 distribution 141 intensity 183 surface crazing 233 swelling 141 switch 282

test fixtures 21 testing conditions 69 textiles 1 Tg 109 TGA 213 thennal conductivity 48 thennolneters 112 thickness 156 through-translnission 121 tie molecules 170 tilt angles 21 titning 24 titaniuln dioxide 84 toys 161 TPO 241 transit shelters 271 translucent polymers 124 transmittance 135, 222 transparent polYlners 124

U unsaturations 215, 228 urethane 302 UV absorbers 70, 77, 164, 185, 233 cycling 265 radiation 3 spectra 228 spectroscopy 186 stabilizers 5

V valves 99 viscosity 275 volatility 253

T temperature 4, 8,49,72, 105, 124, 169, 171,309 gradient 141 tensile 233 strength 12,73, 99, 230, 233 terrestrial body 106

W walkways 271 warping 146 water uptake 155 waterborne 296

320 wavelength 3, 135 Weather-Ometer 9, 72, 94, 242 Wien's Law 127 welding 127 window profile 83

X xenon-arc 7, 16, 43 Xenotest 30, 228 X-ray 197

y yellowness index 273 yield point 216 Young's modulus 146, 151

Z zinc stearate 84

Index

Weathering of Plastics: Testing to Mirror Life Performance (Plastics Design Library) - PDF Free Download (2024)